Computers have revolutionized almost every facet of modern society, and as we approach the physical limits of digital electronics, it becomes imperative to investigate alternative computing hardware paradigms to enable the next generation of faster and more energy-efficient computers. This thesis embarks on building the foundation for a new kind of computer, based on ultrafast nonlinear photonics, aiming to overcome some of the limitations plaguing current computers. In particular, we primarily focus on the clock rate, which has stagnated at \u223c5 GHz for conventional microprocessors over the past two decades.
\r\n\r\nWe begin by identifying single nonlinear devices in lithium niobate nanophotonics that can act as essential building blocks for computers, showing a variety of nonlinear functions with operational speeds > 13 THz for artificial intelligence computing workloads. Then, we progress to small-scale photonic computing circuits combining both strong nonlinearity and memory feedback in a physical reservoir computer for temporal information processing with \u223c10 GHz clock rates. Additionally, we explore unconventional computer architectures such as Cellular Automata, which reveals key system-level considerations that maximize the benefits of ultrafast nonlinear photonics in large-scale computers. This culminates in the demonstration of truly end-to-end and all-optical computing with > 100 GHz clock rates, which represents over an order-of-magnitude advancement compared to existing electronic computers. Finally, we prove mathematically how coupled nonlinear optical resonators are Turing-complete computers.
\r\n\r\nOverall, this work builds on the recent advances in nonlinear photonics and highlights a path for a new class of ultrafast photonic computers that can surpass the clock rate and latency limits of electronic computers, hence enabling nascent applications requiring real-time control or information processing at picosecond timescales.
", "doi": "10.7907/g8re-9a27", "publication_date": "2025", "thesis_type": "phd", "thesis_year": "2025" }, { "id": "thesis:16496", "collection": "thesis", "collection_id": "16496", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:06042024-004345614", "type": "thesis", "title": "Combining Sources and Leveraging Contexts", "author": [ { "family_name": "Mazaheri", "given_name": "Bijan Henrik Socrates", "orcid": "0000-0001-9690-8686", "clpid": "Mazaheri-Bijan-Henrik-Socrates" } ], "thesis_advisor": [ { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Schulman", "given_name": "Leonard J.", "orcid": "0000-0001-9901-2797", "clpid": "Schulman-L-J" } ], "thesis_committee": [ { "family_name": "Eberhardt", "given_name": "Frederick D.", "clpid": "Eberhardt-F-D" }, { "family_name": "Janzing", "given_name": "Dominik", "orcid": "0000-0002-2126-5740", "clpid": "Janzing-Dominik" }, { "family_name": "Rabani", "given_name": "Yuval", "orcid": "0000-0001-7772-2544", "clpid": "Rabani-Yuval" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Schulman", "given_name": "Leonard J.", "orcid": "0000-0001-9901-2797", "clpid": "Schulman-L-J" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "In this thesis we discuss two levels of knowledge beyond regression and classification. The first involves the identification of exchangeable scenarios or individuals from which causal relationships can be ascertained. We discuss one key difficulty of this task, the \"Multi-Source Conundrum,\" which emerges whenever data is merged from multiple sources. This motivates the \"Principle of Limited Latent Classes,\" an assumption which allows us to introduce new algorithms for deconfounding and causal structure learning.
\r\n\r\nThe second level of knowledge involves the expansion from contextual exchangeability to contextual synthesis. We will study a paradox of nontransitivity that occurs when combining multiple contexts, as well a demonstrating robustness gains from using context-dependent counterfactuals as training features. Through these points, we present contextual synthesis as a new frontier with promise for advances in out-of-distribution robustness, fairness, and privacy.
", "doi": "10.7907/7t4d-xg91", "publication_date": "2024", "thesis_type": "phd", "thesis_year": "2024" }, { "id": "thesis:16476", "collection": "thesis", "collection_id": "16476", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:06032024-055145735", "primary_object_url": { "basename": "kilic_kordag_thesis_final.pdf", "content": "final", "filesize": 1747501, "license": "other", "mime_type": "application/pdf", "url": "/16476/8/kilic_kordag_thesis_final.pdf", "version": "v12.0.0" }, "type": "thesis", "title": "On the Complexity of Neural Network Representations", "author": [ { "family_name": "Kili\u00e7", "given_name": "Korda\u011f Mehmet", "orcid": "0009-0005-6321-7113", "clpid": "Kili\u00e7-Korda\u011f-Mehmet" } ], "thesis_advisor": [ { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" } ], "thesis_committee": [ { "family_name": "Hassibi", "given_name": "Babak", "orcid": "0000-0002-1375-5838", "clpid": "Hassibi-B" }, { "family_name": "Winfree", "given_name": "Erik", "orcid": "0000-0002-5899-7523", "clpid": "Winfree-E" }, { "family_name": "Umans", "given_name": "Christopher M.", "orcid": "0000-0002-6390-9401", "clpid": "Umans-C-M" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "The evolution of the human brain was one of the milestones in the history of information after the emergence of life. The underlying biological, chemical, and physical processes of the brain have amazed scientists for a long time. It is still a mystery how the human brain computes a simple arithmetical operation like 2 + 2 = 4. This enigma has spurred investigations into understanding the intrinsic architecture of the brain.
\r\n \r\nThis thesis delves into two primary models for brain architecture: Feedforward Neural Networks and Nearest Neighbor (NN) Representations. Both models are treated under the hypothesis that our brain does not work with \"large\" numbers and expressive power is derived from connectivity. Thus, when examining a network or, more precisely, a single neuron model, we strive to minimize the bit resolution of weights, potentially increasing depth or circuit complexity.
\r\n \r\nFor the NN representations, the memory is defined by a set of vectors in R\u207f (that we call anchors), computation is performed by convergence from an input vector to a nearest neighbor anchor, and the output is a label associated with an anchor. Limited bit resolution in the anchor entries may result in an increase of the size of the NN representation.
\r\n\r\nIn the digital age, computers universally employ the binary numeral system, ensuring the enduring relevance of Boolean functions. This study specifically explores the trade-off between resolution and size for the computation models for Boolean functions. It is established that \"low resolution\" models may require a polynomial or even an exponential increase in the size complexity of the \"high resolution\" model, potentially making the practical implementation infeasible. Building upon prior research, our goal is to optimize these blow-ups by narrowing the gaps between theoretical upper and lower bounds under various constraints. Additionally, we aim to establish connections between NN representations and neural network models by providing explicit NN representations for well-known Boolean functions in Circuit Complexity Theory.
", "doi": "10.7907/trse-ff38", "publication_date": "2024", "thesis_type": "phd", "thesis_year": "2024" }, { "id": "thesis:16496", "collection": "thesis", "collection_id": "16496", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:06042024-004345614", "type": "thesis", "title": "Combining Sources and Leveraging Contexts", "author": [ { "family_name": "Mazaheri", "given_name": "Bijan Henrik Socrates", "orcid": "0000-0001-9690-8686", "clpid": "Mazaheri-Bijan-Henrik-Socrates" } ], "thesis_advisor": [ { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Schulman", "given_name": "Leonard J.", "orcid": "0000-0001-9901-2797", "clpid": "Schulman-L-J" } ], "thesis_committee": [ { "family_name": "Eberhardt", "given_name": "Frederick D.", "clpid": "Eberhardt-F-D" }, { "family_name": "Janzing", "given_name": "Dominik", "orcid": "0000-0002-2126-5740", "clpid": "Janzing-Dominik" }, { "family_name": "Rabani", "given_name": "Yuval", "orcid": "0000-0001-7772-2544", "clpid": "Rabani-Yuval" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Schulman", "given_name": "Leonard J.", "orcid": "0000-0001-9901-2797", "clpid": "Schulman-L-J" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "In this thesis we discuss two levels of knowledge beyond regression and classification. The first involves the identification of exchangeable scenarios or individuals from which causal relationships can be ascertained. We discuss one key difficulty of this task, the \"Multi-Source Conundrum,\" which emerges whenever data is merged from multiple sources. This motivates the \"Principle of Limited Latent Classes,\" an assumption which allows us to introduce new algorithms for deconfounding and causal structure learning.
\r\n\r\nThe second level of knowledge involves the expansion from contextual exchangeability to contextual synthesis. We will study a paradox of nontransitivity that occurs when combining multiple contexts, as well a demonstrating robustness gains from using context-dependent counterfactuals as training features. Through these points, we present contextual synthesis as a new frontier with promise for advances in out-of-distribution robustness, fairness, and privacy.
", "doi": "10.7907/7t4d-xg91", "publication_date": "2024", "thesis_type": "phd", "thesis_year": "2024" }, { "id": "thesis:15004", "collection": "thesis", "collection_id": "15004", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:08182022-183119533", "primary_object_url": { "basename": "NianGuo_2023_thesis.pdf", "content": "final", "filesize": 2085059, "license": "other", "mime_type": "application/pdf", "url": "/15004/1/NianGuo_2023_thesis.pdf", "version": "v6.0.0" }, "type": "thesis", "title": "Causal Sampling, Compressing, and Channel Coding of Streaming Data", "author": [ { "family_name": "Guo", "given_name": "Nian", "orcid": "0000-0003-4490-328X", "clpid": "Guo-Nian" } ], "thesis_advisor": [ { "family_name": "Kostina", "given_name": "Victoria", "orcid": "0000-0002-2406-7440", "clpid": "Kostina-V" } ], "thesis_committee": [ { "family_name": "Effros", "given_name": "Michelle", "orcid": "0000-0003-3757-0675", "clpid": "Effros-M" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Hassibi", "given_name": "Babak", "orcid": "0000-0002-1375-5838", "clpid": "Hassibi-B" }, { "family_name": "Kostina", "given_name": "Victoria", "orcid": "0000-0002-2406-7440", "clpid": "Kostina-V" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "With the emergence of the Internet of Things, communication systems, such as those employed in distributed control and tracking scenarios, are becoming increasingly dynamic, interactive, and delay-sensitive. The data in such real-time systems arrive at the encoder progressively in a streaming fashion. An intriguing question is: what codes can transmit streaming data with both high reliability and low latency? Classical non-causal (block) encoding schemes can transmit data reliably but under the assumption that the encoder knows the entire data block before the transmission. While this is a realistic assumption in delay-tolerant systems, it is ill-suited to real-time systems due to the delay introduced by collecting data into a block. This thesis studies causal encoding: the encoder transmits information based on the causally received data while the data is still streaming in and immediately incorporates the newly received data into a continuing transmission on the fly.
\r\n\r\nThis thesis investigates causal encoding of streaming data in three scenarios: causal sampling, causal lossy compressing, and causal joint source-channel coding (JSCC). In the causal sampling scenario, a sampler observes a continuous-time source process and causally decides when to transmit real-valued samples of it under a constraint on the average number of samples per second; an estimator uses the causally received samples to approximate the source process in real time. We propose a causal sampling policy that achieves the best tradeoff between the sampling frequency and the end-to-end real-time estimation distortion for a class of continuous Markov processes. In the causal lossy compressing scenario, the sampling frequency constraint in the causal sampling scenario is replaced by a rate constraint on the average number of bits per second. We propose a causal code that achieves the best causal distortion-rate tradeoff for the same class of processes. In the causal JSCC scenario, the noiseless channel and the continuous-time process in the previous scenarios are replaced by a discrete memoryless channel with feedback and a sequence of streaming symbols, respectively. We propose a causal joint sourcechannel code that achieves the maximum exponentially decaying rate of the error probability compatible with a given rate. Remarkably, the fundamental limits in the causal lossy compressing and the causal JSCC scenarios achieved by our causal codes are no worse than those achieved by the best non-causal codes. In addition to deriving the fundamental limits and presenting the causal codes that achieve the limits, we also show that our codes apply to control systems, are resilient to system deficiencies such as channel delay and noise, and have low complexities.
", "doi": "10.7907/n201-ca08", "publication_date": "2023", "thesis_type": "phd", "thesis_year": "2023" }, { "id": "thesis:15253", "collection": "thesis", "collection_id": "15253", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:06012023-003751035", "type": "thesis", "title": "Non-Asymptotic Analysis of Single-Receiver Channels with Limited Feedback", "author": [ { "family_name": "Yavas", "given_name": "Recep Can", "orcid": "0000-0002-5640-515X", "clpid": "Yavas-Recep-Can" } ], "thesis_advisor": [ { "family_name": "Effros", "given_name": "Michelle", "orcid": "0000-0003-3757-0675", "clpid": "Effros-M" }, { "family_name": "Kostina", "given_name": "Victoria", "orcid": "0000-0002-2406-7440", "clpid": "Kostina-V" } ], "thesis_committee": [ { "family_name": "Hassibi", "given_name": "Babak", "orcid": "0000-0002-1375-5838", "clpid": "Hassibi-B" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Kostina", "given_name": "Victoria", "orcid": "0000-0002-2406-7440", "clpid": "Kostina-V" }, { "family_name": "Effros", "given_name": "Michelle", "orcid": "0000-0003-3757-0675", "clpid": "Effros-M" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "Emerging Internet of Things, machine-type communication, and ultra-reliable low-latency communication in 5G demand codes that operate at short blocklengths, have low error probability and low energy consumption, and can handle the random activity of a large number of communicating devices. Since many of the applications have a single central device, e.g., a base station, that resolves the communication and a varying number of users, these requirements on the code design motivate interest in the non-asymptotic analysis of codes in a variety of single-receiver channels. This thesis investigates three channel coding problems with the goals of understanding the fundamental limits of channel coding under stringent requirements on reliability, delay, and power, and proposes novel coding architectures that employ constrained feedback to attain those limits. In the first part, we consider point-to-point channels without feedback, and analyze the non-asymptotic limits in the moderate deviations regime in probability theory. The moderate deviations regime is suitable for accurately approximating the maximum achievable coding rate in the operational regimes of practical interest because it simultaneously considers high rates and low error probabilities. We propose a new quantity, channel skewness, which governs the fundamental limit at short blocklengths and low error probabilities. Our approximation is the tightest among the state-of-the-art approximations for most error probability and latency constraints of interest. In the second part, we investigate rateless channel coding with limited feedback. Here, rateless means that decoding can occur at multiple decoding times. In our code design, feedback is limited both in frequency and content; it is sparse, meaning that it is available only at a few instants throughout the communication epoch; and it is stop-feedback, meaning that the receiver informs the transmitters only about whether decoding has occurred rather than what symbols it has received. Our results demonstrate that sporadically sending a few bits is almost as efficient as sending feedback at every time instant. In the third part, we focus on rateless random access channel codes, where the number of active transmitters is unknown to both the transmitters and the receiver. Our rateless code design that reserves a decoding time for each possible number of active transmitters achieves the same first two terms in the asymptotic expansion of the achievable rate as codes where the transmitter activity is known a priori. This means that, remarkably, the random transmitter activity has almost no effect on achievable rates.
\r\n\r\nTo obtain tight channel coding bounds, we analyze some non-asymptotic and asymptotic state-of-the-art bounds on the probability of the sum of independent and identical random variables, whose applications extend to source coding, hypothesis testing, and many others. In the scenarios where these tools are not directly applicable such as for the Gaussian channel, we propose new techniques to overcome that difficulty.
", "doi": "10.7907/a9e9-he45", "publication_date": "2023", "thesis_type": "phd", "thesis_year": "2023" }, { "id": "thesis:14949", "collection": "thesis", "collection_id": "14949", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:06062022-043503154", "primary_object_url": { "basename": "thesis_jcm.pdf", "content": "final", "filesize": 1007053, "license": "cc_by_nc", "mime_type": "application/pdf", "url": "/14949/1/thesis_jcm.pdf", "version": "v5.0.0" }, "type": "thesis", "title": "Combinatorial and Algebraic Propeties of Nonnegative Matrices", "author": [ { "family_name": "Mehta", "given_name": "Jenish Chetan", "clpid": "Mehta-Jenish-Chetan" } ], "thesis_advisor": [ { "family_name": "Schulman", "given_name": "Leonard J.", "orcid": "0000-0001-9901-2797", "clpid": "Schulman-L-J" }, { "family_name": "Vidick", "given_name": "Thomas G.", "orcid": "0000-0002-6405-365X", "clpid": "Vidick-T" } ], "thesis_committee": [ { "family_name": "Umans", "given_name": "Christopher M.", "clpid": "Umans-C-M" }, { "family_name": "Schulman", "given_name": "Leonard J.", "orcid": "0000-0001-9901-2797", "clpid": "Schulman-L-J" }, { "family_name": "Vidick", "given_name": "Thomas G.", "orcid": "0000-0002-6405-365X", "clpid": "Vidick-T" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "We study the combinatorial and algebraic properties of Nonnegative Matrices. Our results are divided into three different categories.
\r\n\r\n1. We show the first quantitative generalization of the 100 year-old Perron-Frobenius theorem, a fundamental theorem which has been used within diverse areas of mathematics. The Perron-Frobenius theorem shows that any irreducible nonnegative matrix R will have a largest positive eigenvalue r, and every other eigenvalue \u03bb is such that Re\u03bb < R and |\u03bb| \u2264 r. We capture the notion of irreducibility through the widely studied notion of edge expansion \u03c6 of R which intuitively measures how well-connected the underlying digraph of R is, and show a quantitative relation between the spectral gap \u0394 = 1-Re\u03bb/r (where \u03bb \u2260 r has the largest real part) of R to the edge expansion \u03c6 as follows.
\r\n\r\n(1/15) \u2022 [(\u0394(R))/n] \u2264 \u03c6(R) \u2264 \u221a[2 \u2022 \u0394(R)].
\r\n\r\nThis also provides a more general result than the Cheeger-Buser inequalities since it applies to any nonnegative matrix.
\r\n\r\n2. We study constructions of specific nonsymmetric matrices (or nonreversible Markov Chains) that have small edge expansion but large spectral gap, taking us in a direction more novel and unexplored than studying symmetric matrices with constant edge expansion that have been extensively studied. We first analyze some known but less studied Markov Chains, and then provide a novel construction of a nonreversible chain for which
\r\n\r\n\u03c6(R) \u2264 [(\u0394(R))/\u221an],
\r\n\r\nobtaining a bound exponentially better than known bounds. We also present a candidate construction of matrices for which
\r\n\r\n\u03c6(R) \u2264 2[(\u0394(R))/n]
\r\n\r\nwhich is the most beautiful contribution of this thesis. We believe these matrices have properties remarkable enough to deserve study in their own right.
\r\n\r\n3. We connect the edge expansion and spectral gap to other combinatorial properties of nonsymmetric matrices. The most well-studied property is mixing time, and we provide elementary proofs of the relation between mixing time and the edge expansion, and also other bounds relating the mixing time of a nonreversible chain to the spectral gap and to its additive symmetrization. Further, we provide a unified view of the notion of capacity and normalized capacity, and show the monotonicity of capacity of nonreversible chains amongst other results for nonsymmetric matrices. We finally discuss and prove interesting lemmas about different notions of expansion and show the first results for tensor walks or nonnegative tensors.
", "doi": "10.7907/3vxb-6778", "publication_date": "2022", "thesis_type": "phd", "thesis_year": "2022" }, { "id": "thesis:14949", "collection": "thesis", "collection_id": "14949", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:06062022-043503154", "primary_object_url": { "basename": "thesis_jcm.pdf", "content": "final", "filesize": 1007053, "license": "cc_by_nc", "mime_type": "application/pdf", "url": "/14949/1/thesis_jcm.pdf", "version": "v5.0.0" }, "type": "thesis", "title": "Combinatorial and Algebraic Propeties of Nonnegative Matrices", "author": [ { "family_name": "Mehta", "given_name": "Jenish Chetan", "clpid": "Mehta-Jenish-Chetan" } ], "thesis_advisor": [ { "family_name": "Schulman", "given_name": "Leonard J.", "orcid": "0000-0001-9901-2797", "clpid": "Schulman-L-J" }, { "family_name": "Vidick", "given_name": "Thomas G.", "orcid": "0000-0002-6405-365X", "clpid": "Vidick-T" } ], "thesis_committee": [ { "family_name": "Umans", "given_name": "Christopher M.", "clpid": "Umans-C-M" }, { "family_name": "Schulman", "given_name": "Leonard J.", "orcid": "0000-0001-9901-2797", "clpid": "Schulman-L-J" }, { "family_name": "Vidick", "given_name": "Thomas G.", "orcid": "0000-0002-6405-365X", "clpid": "Vidick-T" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "We study the combinatorial and algebraic properties of Nonnegative Matrices. Our results are divided into three different categories.
\r\n\r\n1. We show the first quantitative generalization of the 100 year-old Perron-Frobenius theorem, a fundamental theorem which has been used within diverse areas of mathematics. The Perron-Frobenius theorem shows that any irreducible nonnegative matrix R will have a largest positive eigenvalue r, and every other eigenvalue \u03bb is such that Re\u03bb < R and |\u03bb| \u2264 r. We capture the notion of irreducibility through the widely studied notion of edge expansion \u03c6 of R which intuitively measures how well-connected the underlying digraph of R is, and show a quantitative relation between the spectral gap \u0394 = 1-Re\u03bb/r (where \u03bb \u2260 r has the largest real part) of R to the edge expansion \u03c6 as follows.
\r\n\r\n(1/15) \u2022 [(\u0394(R))/n] \u2264 \u03c6(R) \u2264 \u221a[2 \u2022 \u0394(R)].
\r\n\r\nThis also provides a more general result than the Cheeger-Buser inequalities since it applies to any nonnegative matrix.
\r\n\r\n2. We study constructions of specific nonsymmetric matrices (or nonreversible Markov Chains) that have small edge expansion but large spectral gap, taking us in a direction more novel and unexplored than studying symmetric matrices with constant edge expansion that have been extensively studied. We first analyze some known but less studied Markov Chains, and then provide a novel construction of a nonreversible chain for which
\r\n\r\n\u03c6(R) \u2264 [(\u0394(R))/\u221an],
\r\n\r\nobtaining a bound exponentially better than known bounds. We also present a candidate construction of matrices for which
\r\n\r\n\u03c6(R) \u2264 2[(\u0394(R))/n]
\r\n\r\nwhich is the most beautiful contribution of this thesis. We believe these matrices have properties remarkable enough to deserve study in their own right.
\r\n\r\n3. We connect the edge expansion and spectral gap to other combinatorial properties of nonsymmetric matrices. The most well-studied property is mixing time, and we provide elementary proofs of the relation between mixing time and the edge expansion, and also other bounds relating the mixing time of a nonreversible chain to the spectral gap and to its additive symmetrization. Further, we provide a unified view of the notion of capacity and normalized capacity, and show the monotonicity of capacity of nonreversible chains amongst other results for nonsymmetric matrices. We finally discuss and prove interesting lemmas about different notions of expansion and show the first results for tensor walks or nonnegative tensors.
", "doi": "10.7907/3vxb-6778", "publication_date": "2022", "thesis_type": "phd", "thesis_year": "2022" }, { "id": "thesis:14951", "collection": "thesis", "collection_id": "14951", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:06062022-185003284", "primary_object_url": { "basename": "Caltech_Thesis_Jin Sima.pdf", "content": "final", "filesize": 1460633, "license": "other", "mime_type": "application/pdf", "url": "/14951/1/Caltech_Thesis_Jin Sima.pdf", "version": "v4.0.0" }, "type": "thesis", "title": "Correcting Errors in DNA Storage", "author": [ { "family_name": "Sima", "given_name": "Jin", "orcid": "0000-0003-4588-9790", "clpid": "Sima-Jin" } ], "thesis_advisor": [ { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" } ], "thesis_committee": [ { "family_name": "Kostina", "given_name": "Victoria", "orcid": "0000-0002-2406-7440", "clpid": "Kostina-V" }, { "family_name": "Hassibi", "given_name": "Babak", "orcid": "0000-0002-1375-5838", "clpid": "Hassibi-B" }, { "family_name": "Raviv", "given_name": "Netanel", "orcid": "0000-0002-1686-1994", "clpid": "Raviv-N" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "DNA-based storage has potentially unprecedented advantages of high information density and long duration, and is one of the promising techniques to meet the ever-growing demands to keep data in the future. As noise and errors are present in almost every procedure during reading, writing, and storing of information in DNA storage systems, error correction is inevitable to guarantee reliable data storage in DNA. Moreover, it is often required that error correction is done in an efficient manner to reduce the cost and time needed for reading and writing data. Due to the technology constraints and physical limitations, error correction in DNA-based storage poses the following challenges that differ from those in traditional digital data transmission and storage systems.
\r\n\r\n1. A combination of deletion, insertion, and substitution errors present. The goal is to construct efficient codes correcting these errors. While substitution errors are special cases of deletion and insertion errors, and are well studied under the current theory and practice frameworks, deletion and insertion errors are much more difficult to deal with, and less understanding was gained for deletion and insertion errors.
\r\n\r\n2. Error correction is over an unordered set of strings, rather than over a single string, which can be regarded as a set of ordered strings. The latter, which includes the above deletion/insertion coding problem, is commonly studied for current digital communication and storage systems. Our goal is to extend the deletion/insertion correction capability for a single string to a set of unordered strings.
\r\n\r\n3. The decoder observes multiple noisy copies of every coded string. The problem is to deduce a set of strings (or a single string) from a collection of their noisy samples, also studied as the population recovery (or trace reconstruction for a single string) problem. The problem is well answered with substitution errors only and becomes elusive with the introduction of deletion and insertion errors.
\r\n\r\nThis thesis tries to address the above challenges. For the first challenge, we proposed binary codes correct any constant number of deletions and/or insertions with order-wise optimal redundancy, which made a step toward a solution to a longstanding open problem introduced by Levenshtein in 1960s. We also extended it to different settings, in particular, non-binary deletin/insertion correcting codes suitable for DNA storage applications.
\r\n\r\nFor the second challenge, we established lower and upper bounds on the optimal redundancy of codes correcting any number of substitution, deletion, and insertion errors and found that the redundancy needed for coding over an unordered set of strings is order-wise the same as that needed for coding over a ordered set of strings. Using our results for the first challenge, we proposed codes correcting any constant number of deletion/insertion errors with order-wise optimal redundancy under some parametter settings.
\r\n\r\nFor the third challenge, we studied the problem of trace reconstruction, which asks the number of noisy samples needed to reconstruct a single string. While there is a exponential gap between upper and lower bounds on sample complexities in general, we showed that a polynomial number of samples suffice, given a reference string that is within constant edit distance from the target string.
\r\n\r\nApart from dealing with the above challenges, we investigated error correction for multi-head racetrack memory applications. The problem can be considered as correcting any constant number of deletions/insertions in a single string with multiple noisy copies, with the help of coding. Different from the settings we considered above in the trace reconstruction problem, where noisy copies are independent given the target string, in racetrack memory, the noisy copies are correlated, and the number of errors is small compared to the trace reconstruction problem. We derived a lower bound on redundancy and proposed a code correcting any number of deletions/insertions with order-wise optimal redundancy.
", "doi": "10.7907/kdph-6z71", "publication_date": "2022", "thesis_type": "phd", "thesis_year": "2022" }, { "id": "thesis:14172", "collection": "thesis", "collection_id": "14172", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:05242021-223430388", "primary_object_url": { "basename": "Thesis_Hikmet.pdf", "content": "final", "filesize": 648670, "license": "other", "mime_type": "application/pdf", "url": "/14172/1/Thesis_Hikmet.pdf", "version": "v4.0.0" }, "type": "thesis", "title": "Linear Codes with Constrained Generator Matrices", "author": [ { "family_name": "Yildiz", "given_name": "Hikmet", "orcid": "0000-0002-0891-3352", "clpid": "Yildiz-Hikmet" } ], "thesis_advisor": [ { "family_name": "Hassibi", "given_name": "Babak", "orcid": "0000-0002-1375-5838", "clpid": "Hassibi-B" } ], "thesis_committee": [ { "family_name": "Kostina", "given_name": "Victoria", "orcid": "0000-0002-2406-7440", "clpid": "Kostina-V" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Umans", "given_name": "Christopher M.", "clpid": "Umans-C-M" }, { "family_name": "Hassibi", "given_name": "Babak", "orcid": "0000-0002-1375-5838", "clpid": "Hassibi-B" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "Designing good error correcting codes whose generator matrix has a support constraint, i.e., one for which only certain entries of the generator matrix are allowed to be nonzero, has found many recent applications, including in distributed coding and storage, linear network coding, multiple access networks, and weakly secure data exchange. The dual problem, where the parity check matrix has a support constraint, comes up in the design of locally repairable codes. The central problem here is to design codes with the largest possible minimum distance, subject to the given support constraint on the generator matrix. When the distance metric is the Hamming distance, the codes of interest are Reed-Solomon codes, for which case, the problem was formulated as the \"GM-MDS conjecture.\" In the rank metric case, the same problem can be considered for Gabidulin codes. This thesis provides solutions to these problems and discusses the remaining open problems.
", "doi": "10.7907/qz6m-wp22", "publication_date": "2021", "thesis_type": "phd", "thesis_year": "2021" }, { "id": "thesis:13687", "collection": "thesis", "collection_id": "13687", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:04292020-052418975", "type": "thesis", "title": "Formal Design and Analysis for DNA Implementations of Chemical Reaction Networks", "author": [ { "family_name": "Johnson", "given_name": "Robert Francis", "orcid": "0000-0002-5340-8347", "clpid": "Johnson-Robert-Francis" } ], "thesis_advisor": [ { "family_name": "Winfree", "given_name": "Erik", "orcid": "0000-0002-5899-7523", "clpid": "Winfree-E" }, { "family_name": "Qian", "given_name": "Lulu", "orcid": "0000-0003-4115-2409", "clpid": "Qian-Lulu" } ], "thesis_committee": [ { "family_name": "Murray", "given_name": "Richard M.", "orcid": "0000-0002-5785-7481", "clpid": "Murray-R-M" }, { "family_name": "Pierce", "given_name": "Niles A.", "orcid": "0000-0003-2367-4406", "clpid": "Pierce-N-A" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Winfree", "given_name": "Erik", "orcid": "0000-0002-5899-7523", "clpid": "Winfree-E" }, { "family_name": "Qian", "given_name": "Lulu", "orcid": "0000-0003-4115-2409", "clpid": "Qian-Lulu" } ], "local_group": [ { "literal": "div_bbe" } ], "abstract": "In molecular programming, the Chemical Reaction Network model is often used to describe systems of interacting molecules. This model can describe either real systems, allowing us to analyze and determine their computational function; or describe hypothetical systems, with known computational function but perhaps no known physical example. One significant breakthrough in the field is that any Chemical Reaction Network can be approximated by a system using DNA Strand Displacement mechanisms. This allows the Chemical Reaction Network model to be treated like a programming language, where programs can be written in the abstract and then compiled into physical molecules. Given a programming language and a proof-of-concept compiler, one would want to take the compiler from the proof-of-concept stage into a more reliable, more systematic, and better understood process. This thesis is made up of my contributions to that effort.
\r\n\r\nFirst, given a programming language and a compiler, it would be useful to formally verify that the compiler is correct. My collaborators, Qing Dong and Erik Winfree, and I defined a Chemical Reaction Network-specific form of bisimulation equivalence, which can compare two such networks and verify that one is (or is not) a correct implementation of the other. For example, the compiler-produced DNA circuit can be verified as an implementation of its abstract program, although this is not the only possible use. After defining this concept of equivalence, we show that it can be checked by algorithm; although various parts of the problem are NP-complete or PSPACE-complete, we give algorithms that meet these lower bounds. We also prove a number of interesting properties of Chemical Reaction Network bisimulation equivalence, including transitivity and modularity properties which are particularly useful for stepwise checking of large systems. We further extend this bisimulation method to linear Polymer Reaction Networks, a strictly more powerful abstraction which has been occasionally used in molecular programming. Again we prove complexity hardness results, which in this case are as expected uncomputable in the general case; however, many practical systems can still be verified, and we give one such example. Finally, we use bisimulation to identify a class of single-locus networks that are practical to implement. Thus we show a method of verification which can simplify use of the above-mentioned compiler by proving general statements of correctness about its results.
\r\n\r\nSecond, given a programming language and a concept of compiling it, it would be useful to optimize the result of the compilation. One particular area of optimization is the number of DNA strands per prepared complex; some experiments suggest that systems with no more than 2 strands per complex are more robust. Lulu Qian and I developed some proposed DNA Strand Displacement schemes for general Chemical Reaction Network implementations with no more than 2 strands per complex, and a number of other desirable properties. Meanwhile, having been shown to be useful for many reasons, the mechanisms of DNA Strand Displacement have recently been formalized, abstracted, and analyzed. I show that this formalization, combined with the bisimulation methods above, can prove various statements about the limits of DNA Strand Displacement systems. For example, a set of desirable conditions including the 2-strand limit cannot be achieved by any general Chemical Reaction Network implementation scheme. I also observe that two of the new schemes we discovered, each meeting all but one condition of the impossible set, were found in the process of coming up with this proof. I thus argue that through formalization of DNA Strand Displacement we can have a more systematic method of finding and designing molecular programs, and of knowing when the programs we want do not exist.
", "doi": "10.7907/a74v-kv80", "publication_date": "2020", "thesis_type": "phd", "thesis_year": "2020" }, { "id": "thesis:11436", "collection": "thesis", "collection_id": "11436", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:04032019-102853075", "type": "thesis", "title": "Decoding the Past", "author": [ { "family_name": "Jain", "given_name": "Siddharth", "orcid": "0000-0002-9164-6119", "clpid": "Jain-Siddharth" } ], "thesis_advisor": [ { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" } ], "thesis_committee": [ { "family_name": "Vaidyanathan", "given_name": "P. P.", "clpid": "Vaidyanathan-P-P" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Hassibi", "given_name": "Babak", "clpid": "Hassibi-B" }, { "family_name": "Winfree", "given_name": "Erik", "clpid": "Winfree-E" }, { "family_name": "Schwartz", "given_name": "Moshe", "clpid": "Schwartz-Moshe" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "The human genome is continuously evolving, hence the sequenced genome is a snapshot in time of this evolving entity. Over time, the genome accumulates mutations that can be associated with different phenotypes - like physical traits, diseases, etc. Underlying mutation accumulation is an evolution channel (the term channel is motivated by the notion of communication channel introduced by Shannon [1] in 1948 and started the area of Information Theory), which is controlled by hereditary, environmental, and stochastic factors. The premise of this thesis is to understand the human genome using information theory framework. In particular, it focuses on: (i) the analysis and characterization of the evolution channel using measures of capacity, expressiveness, evolution distance, and uniqueness of ancestry and uses these insights for (ii) the design of error correcting codes for DNA storage, (iii) inversion symmetry in the genome and (iv) cancer classification.
\r\n\r\nThe mutational events characterizing this evolution channel can be divided into two categories, namely point mutations and duplications. While evolution through point mutations is unconstrained, giving rise to combinatorially many possibilities of what could have happened in the past, evolution through duplications adds constraints limiting the number of those possibilities. Further, more than 50% of the genome has been observed to consist of repeated sequences. We focus on the much constrained form of duplications known as tandem duplications in order to understand the limits of evolution by duplication. Our sequence evolution model consists of a starting sequence called seed and a set of tandem duplication rules. We find limits on the diversity of sequences that can be generated by tandem duplications using measures of capacity and expressiveness. Additionally, we calculate bounds on the duplication distance which is used to measure the timing of generation by these duplications. We also ask questions about the uniqueness of seed for a given sequence and completely characterize the duplication length sets where the seed is unique or non-unique. These insights also led us to design error correcting codes for any number of tandem duplication errors that are useful for DNA-storage based applications. For uniform duplication length and duplication length bounded by 2, our designed codes achieve channel capacity. We also define and measure uncertainty in decoding when the duplication channel is misinformed. Moreover, we add substitutions to our tandem duplication model and calculate sequence generation diversity for a given budget of substitutions.
\r\n\r\nWe also use our duplication model to explain the inversion symmetry observed in the genome of many species. The inversion symmetry is popularly known as the 2nd Chargaff Rule, according to which in a single strand DNA, the frequency of a k-mer is almost the same as the frequency of its reverse complement. The insights gained by these problems led us to investigate the tandem repeat regions in the genome. Tandem repeat regions in the genome can be traced back in time algorithmically to make inference about the effect of the hereditary, environmental and stochastic factors on the mutation rate of the genome. By inferring the evolutionary history of the tandem repeat regions, we show how this knowledge can be used to make predictions about the risk of incurring a mutation based disease, specifically cancer. More precisely, we introduce the concept of mutation profiles that are computed without any comparative analysis, but instead by analyzing the short tandem repeat regions in a single healthy genome and capturing information about the individual's evolution channel. Using gradient boosting on data from more than 5,000 TCGA (The Cancer Genome Atlas) cancer patients, we demonstrate that these mutation profiles can accurately distinguish between patients with various types of cancer. For example, the pairwise validation accuracy of the classifier between PAAD (pancreas) patients and GBM (brain) patients is 93%. Our results show that healthy unaffected cells still contain a cancer-specific signal, which opens the possibility of cancer prediction from a healthy genome.
", "doi": "10.7907/K286-5N63", "publication_date": "2019", "thesis_type": "phd", "thesis_year": "2019" }, { "id": "thesis:10970", "collection": "thesis", "collection_id": "10970", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:05302018-095132389", "primary_object_url": { "basename": "Liu_Chun-Lin_2018.pdf", "content": "final", "filesize": 2762411, "license": "other", "mime_type": "application/pdf", "url": "/10970/1/Liu_Chun-Lin_2018.pdf", "version": "v8.0.0" }, "type": "thesis", "title": "Sparse Array Signal Processing: New Array Geometries, Parameter Estimation, and Theoretical Analysis", "author": [ { "family_name": "Liu", "given_name": "Chun-Lin", "orcid": "0000-0003-3135-9684", "clpid": "Liu-Chun-Lin" } ], "thesis_advisor": [ { "family_name": "Vaidyanathan", "given_name": "P. P.", "clpid": "Vaidyanathan-P-P" } ], "thesis_committee": [ { "family_name": "Vaidyanathan", "given_name": "P. P.", "clpid": "Vaidyanathan-P-P" }, { "family_name": "Abu-Mostafa", "given_name": "Yaser S.", "clpid": "Abu-Mostafa-Y-S" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Kostina", "given_name": "Victoria", "clpid": "Kostina-V" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "Array signal processing focuses on an array of sensors receiving the incoming waveforms in the environment, from which source information, such as directions of arrival (DOA), signal power, amplitude, polarization, and velocity, can be estimated. This topic finds ubiquitous applications in radar, astronomy, tomography, imaging, and communications. In these applications, sparse arrays have recently attracted considerable attention, since they are capable of resolving O(N2) uncorrelated source directions with N physical sensors. This is unlike the uniform linear arrays (ULA), which identify at most N-1 uncorrelated sources with N sensors. These sparse arrays include minimum redundancy arrays (MRA), nested arrays, and coprime arrays. All these arrays have an O(N2)-long central ULA segment in the difference coarray, which is defined as the set of differences between sensor locations. This O(N2) property makes it possible to resolve O(N2) uncorrelated sources, using only N physical sensors.
\r\n\r\nThe main contribution of this thesis is to provide a new direction for array geometry and performance analysis of sparse arrays in the presence of nonidealities. The first part of this thesis focuses on designing novel array geometries that are robust to effects of mutual coupling. It is known that, mutual coupling between sensors has an adverse effect on the estimation of DOA. While there are methods to counteract this through appropriate modeling and calibration, they are usually computationally expensive, and sensitive to model mismatch. On the other hand, sparse arrays, such as MRA, nested arrays, and coprime arrays, have reduced mutual coupling compared to ULA, but all of these have their own disadvantages. This thesis introduces a new array called the super nested array, which has many of the good properties of the nested array, and at the same time achieves reduced mutual coupling. Many theoretical properties are proved and simulations are included to demonstrate the superior performance of super nested arrays in the presence of mutual coupling.
\r\n\r\nTwo-dimensional planar sparse arrays with large difference coarrays have also been known for a long time. These include billboard arrays, open box arrays (OBA), and 2D nested arrays. However, all of them have considerable mutual coupling. This thesis proposes new planar sparse arrays with the same large difference coarrays as the OBA, but with reduced mutual coupling. The new arrays include half open box arrays (HOBA), half open box arrays with two layers (HOBA-2), and hourglass arrays. Among these, simulations show that hourglass arrays have the best estimation performance in presence of mutual coupling.
\r\n\r\nThe second part of this thesis analyzes the performance of sparse arrays from a theoretical perspective. We first study the Cram\u00e9r-Rao bound (CRB) for sparse arrays, which poses a lower bound on the variances of unbiased DOA estimators. While there exist landmark papers on the study of the CRB in the context of array processing, the closed-form expressions available in the literature are not applicable in the context of sparse arrays for which the number of identifiable sources exceeds the number of sensors. This thesis derives a new expression for the CRB to fill this gap. Based on the proposed CRB expression, it is possible to prove the previously known experimental observation that, when there are more sources than sensors, the CRB stagnates to a constant value as the SNR tends to infinity. It is also possible to precisely specify the relation between the number of sensors and the number of uncorrelated sources such that these sources could be resolved.
\r\n\r\nRecently, it has been shown that correlation subspaces, which reveal the structure of the covariance matrix, help to improve some existing DOA estimators. However, the bases, the dimension, and other theoretical properties of correlation subspaces remain to be investigated. This thesis proposes generalized correlation subspaces in one and multiple dimensions. This leads to new insights into correlation subspaces and DOA estimation with prior knowledge. First, it is shown that the bases and the dimension of correlation subspaces are fundamentally related to difference coarrays, which were previously found to be important in the study of sparse arrays. Furthermore, generalized correlation subspaces can handle certain forms of prior knowledge about source directions. These results allow one to derive a broad class of DOA estimators with improved performance.
\r\n\r\nIt is empirically known that the coarray structure is susceptible to sensor failures, and the reliability of sparse arrays remains a significant but challenging topic for investigation. This thesis advances a general theory for quantifying such robustness, by studying the effect of sensor failure on the difference coarray. We first present the (k-)essentialness property, which characterizes the combinations of the faulty sensors that shrink the difference coarray. Based on this, the notion of (k-)fragility is proposed to quantify the reliability of sparse arrays with faulty sensors, along with comprehensive studies of their properties. These novel concepts provide quite a few insights into the interplay between the array geometry and its robustness. For instance, for the same number of sensors, it can be proved that ULA is more robust than the coprime array, and the coprime array is more robust than the nested array. Rigorous development of these ideas leads to expressions for the probability of coarray failure, as a function of the probability of sensor failure.
\r\n\r\nThe thesis concludes with some remarks on future directions and open problems.
", "doi": "10.7907/NSTQ-SD57", "publication_date": "2018", "thesis_type": "phd", "thesis_year": "2018" }, { "id": "thesis:11029", "collection": "thesis", "collection_id": "11029", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:06062018-132643508", "primary_object_url": { "basename": "tenneti-srikanth-venkata-thesis-V4.pdf", "content": "final", "filesize": 14012477, "license": "other", "mime_type": "application/pdf", "url": "/11029/14/tenneti-srikanth-venkata-thesis-V4.pdf", "version": "v5.0.0" }, "type": "thesis", "title": "The Nested Periodic Subspaces: Extensions of Ramanujan Sums for Period Estimation", "author": [ { "family_name": "Tenneti", "given_name": "Srikanth Venkata", "orcid": "0000-0002-5415-3681", "clpid": "Tenneti-Srikanth-Venkata" } ], "thesis_advisor": [ { "family_name": "Vaidyanathan", "given_name": "P. P.", "clpid": "Vaidyanathan-P-P" } ], "thesis_committee": [ { "family_name": "Vaidyanathan", "given_name": "P. P.", "clpid": "Vaidyanathan-P-P" }, { "family_name": "Kostina", "given_name": "Victoria", "clpid": "Kostina-V" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Abu-Mostafa", "given_name": "Yaser S.", "clpid": "Abu-Mostafa-Y-S" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "In the year 1918, the Indian mathematician Srinivasa Ramanujan proposed a set of sequences called Ramanujan Sums as bases to expand arithmetic functions in number theory. Today, exactly a 100 years later, we will show that these sequences re-emerge as exciting tools in a completely different context: For the extraction of periodic patterns in data. Combined with the state-of-the-art techniques of DSP, Ramanujan Sums can be used as the starting point for developing powerful algorithms for periodicity applications.
\r\n\r\nThe primary inspiration for this thesis comes from a recent extension of Ramanujan sums to subspaces known as the Ramanujan subspaces. These subspaces were designed to span any sequence with integer periodicity, and have many interesting properties. Starting with Ramanujan subspaces, this thesis first develops an entire family of such subspace representations for periodic sequences. This family, called Nested Periodic Subspaces due to their unique structure, turns out to be the least redundant sets of subspaces that can span periodic sequences.
\r\n\r\nThree classes of new algorithms are proposed using the Nested Periodic Subspaces: dictionaries, filter banks, and eigen-space methods based on the auto-correlation matrix of the signal. It will be shown that these methods are especially advantageous to use when the data-length is short, or when the signal is a mixture of multiple hidden periods. The dictionary techniques were inspired by recent advances in sparsity based compressed sensing. Apart from the l1 norm based convex programs currently used in other applications, our dictionaries can admit l2 norm formulations that have linear and closed form solutions, even when the systems is under-determined. A new filter bank is also proposed using the Ramanujan sums. This, named the Ramanujan Filter Bank, can accurately track the instantaneous period for signals that exhibit time varying periodic nature. The filters in the Ramanujan Filter Bank have simple integer valued coefficients, and directly tile the period vs time plane, unlike classical STFT (Short Time Fourier Transform) and wavelets, which tile the time-frequency plane. The third family of techniques developed here are a generalization of the classic MUSIC (MUltiple SIgnal Classification) algorithm for periodic signals. MUSIC is one of the most popular techniques today for line spectral estimation. However, periodic signals are not just any unstructured line spectral signals. There is a nice harmonic spacing between the lines which is not exploited by plain MUSIC. We will show that one can design much more accurate adaptations of MUSIC using Nested Periodic Subspaces. Compared to prior variants of MUSIC for the periodicity problem, our approach is much faster and yields much more accurate results for signals with integer periods. This work is also the first extension of MUSIC that uses simple integer valued basis vectors instead of using traditional complex-exponentials to span the signal subspace. The advantages of the new methods are demonstrated both on simulations, as well as real world applications such as DNA micro-satellites, protein repeats and absence seizures.
\r\n\r\nApart from practical contributions, the theory of Nested Periodic Subspaces offers answers to a number of fundamental questions that were previously unanswered. For example, what is the minimum contiguous data-length needed to be able to identify the period of a signal unambiguously? Notice that the answer we seek is a fundamental identifiability bound independent of any particular period estimation technique. Surprisingly, this basic question has never been answered before. In this thesis, we will derive precise expressions for the minimum necessary and sufficient datalengths for this question. We also extend these bounds to the context of mixtures of periodic signals. Once again, even though mixtures of periodic signals often occur in many applications, aspects such as the unique identifiability of the component periods were never rigorously analyzed before. We will present such an analysis as well.
\r\n\r\nWhile the above question deals with the minimum contiguous datalength required for period estimation, one may ask a slightly different question: If we are allowed to pick the samples of a signal in a non-contiguous fashion, how should we pick them so that we can estimate the period using the least number of samples? This question will be shown to be quite difficult to answer in general. In this thesis, we analyze a smaller case in this regard, namely, that of resolving between two periods. It will be shown that the analysis is quite involved even in this case, and the optimal sampling pattern takes an interesting form of sparsely located bunches. This result can also be extended to the case of multi-dimensional periodic signals.
\r\n\r\nWe very briefly address multi-dimensional periodicity in this thesis. Most prior DSP literature on multi-dimensional discrete time periodic signals assumes the period to be parallelepipeds. But as shown by the artist M. C. Escher, one can tile the space using a much more diverse variety of shapes. Is it always possible to account for such other periodic shapes using the traditional notion of parallelepiped periods? An interesting analysis in this regard is presented towards the end of the thesis.
\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n", "doi": "10.7907/1n4t-5876", "publication_date": "2018", "thesis_type": "phd", "thesis_year": "2018" }, { "id": "thesis:10292", "collection": "thesis", "collection_id": "10292", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:06052017-221842127", "type": "thesis", "title": "Boosting Boosting", "author": [ { "family_name": "Appel", "given_name": "Ron", "clpid": "Appel-Ron" } ], "thesis_advisor": [ { "family_name": "Perona", "given_name": "Pietro", "clpid": "Perona-P" } ], "thesis_committee": [ { "family_name": "Perona", "given_name": "Pietro", "clpid": "Perona-P" }, { "family_name": "Abu-Mostafa", "given_name": "Yaser S.", "clpid": "Abu-Mostafa-Y-S" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Hassibi", "given_name": "Babak", "clpid": "Hassibi-B" }, { "family_name": "Yue", "given_name": "Yisong", "clpid": "Yue-Yisong" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "Machine learning is becoming prevalent in all aspects of our lives. For some applications, there is a need for simple but accurate white-box systems that are able to train efficiently and with little data.
\r\n\r\n\"Boosting\" is an intuitive method, combining many simple (possibly inaccurate) predictors to form a powerful, accurate classifier. Boosted classifiers are intuitive, easy to use, and exhibit the fastest speeds at test-time when implemented as a cascade. However, they have a few drawbacks: training decision trees is a relatively slow procedure, and from a theoretical standpoint, no simple unified framework for cost-sensitive multi-class boosting exists. Furthermore, (axis-aligned) decision trees may be inadequate in some situations, thereby stalling training; and even in cases where they are sufficiently useful, they don't capture the intrinsic nature of the data, as they tend to form boundaries that overfit.
\r\n\r\nMy thesis focuses on remedying these three drawbacks of boosting.\r\nCh.III outlines a method (called QuickBoost) that trains identical classifiers at an order of magnitude faster than before, based on a proof of a bound. In Ch.IV, a unified framework for cost-sensitive multi-class boosting (called REBEL) is proposed, both advancing theory and demonstrating empirical gains. Finally, Ch.V describes a novel family of weak learners (called Localized Similarities) that guarantee theoretical bounds and outperform decision trees and Neural Nets (as well as several other commonly used classification methods) on a range of datasets.
\r\n\r\nThe culmination of my work is an easy-to-use, fast-training, cost-sensitive multi-class boosting framework whose functionality is interpretable (since each weak learner is a simple comparison of similarity), and whose performance is better than Neural Networks and other competing methods. It is the tool that everyone should have in their toolbox and the first one they try.
", "doi": "10.7907/Z9N29V0J", "publication_date": "2017", "thesis_type": "phd", "thesis_year": "2017" }, { "id": "thesis:10269", "collection": "thesis", "collection_id": "10269", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:06042017-212503971", "primary_object_url": { "basename": "Huang_Wentao_2017.pdf", "content": "final", "filesize": 3753004, "license": "other", "mime_type": "application/pdf", "url": "/10269/1/Huang_Wentao_2017.pdf", "version": "v4.0.0" }, "type": "thesis", "title": "Coding for Security and Reliability in Distributed Systems", "author": [ { "family_name": "Huang", "given_name": "Wentao", "orcid": "0000-0003-0963-3624", "clpid": "Huang-Wentao" } ], "thesis_advisor": [ { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" } ], "thesis_committee": [ { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Low", "given_name": "Steven H.", "clpid": "Low-S-H" }, { "family_name": "Hassibi", "given_name": "Babak", "clpid": "Hassibi-B" }, { "family_name": "Effros", "given_name": "Michelle", "clpid": "Effros-M" }, { "family_name": "Langberg", "given_name": "Michael", "clpid": "Langberg-M" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "This dissertation studies the use of coding techniques to improve the reliability and security of distributed systems. The first three parts focus on distributed storage systems, and study schemes that encode a message into n shares, assigned to n nodes, such that any n - r nodes can decode the message (reliability) and any colluding z nodes cannot infer any information about the message (security). The objective is to optimize the computational, implementation, communication and access complexity of the schemes during the process of encoding, decoding and repair. These are the key metrics of the schemes so that when they are applied in practical distributed storage systems, the systems are not only reliable and secure, but also fast and cost-effective.
\r\n\r\nSchemes with highly efficient computation and implementation are studied in Part I. For the practical high rate case of r \u2264 3 and z \u2264 3, we construct schemes that require only r + z XORs to encode and z XORs to decode each message bit, based on practical erasure codes including the B, EVENODD and STAR codes. This encoding and decoding complexity is shown to be optimal. For general r and z, we design schemes over a special ring from Cauchy matrices and Vandermonde matrices. Both schemes can be efficiently encoded and decoded due to the structure of the ring. We also discuss methods to shorten the proposed schemes.
\r\n\r\nPart II studies schemes that are efficient in terms of communication and access complexity. We derive a lower bound on the decoding bandwidth, and design schemes achieving the optimal decoding bandwidth and access. We then design schemes that achieve the optimal bandwidth and access not only for decoding, but also for repair. Furthermore, we present a family of Shamir's schemes with asymptotically optimal decoding bandwidth.
\r\n\r\nPart III studies the problem of secure repair, i.e., reconstructing the share of a (failed) node without leaking any information about the message. We present generic secure repair protocols that can securely repair any linear schemes. We derive a lower bound on the secure repair bandwidth and show that the proposed protocols are essentially optimal in terms of bandwidth.
\r\n\r\nIn the final part of the dissertation, we study the use of coding techniques to improve the reliability and security of network communication.
\r\n\r\nSpecifically, in Part IV we draw connections between several important problems in network coding. We present reductions that map an arbitrary multiple-unicast network coding instance to a unicast secure network coding instance in which at most one link is eavesdropped, or a unicast network error correction instance in which at most one link is erroneous, such that a rate tuple is achievable in the multiple-unicast network coding instance if and only if a corresponding rate is achievable in the unicast secure network coding instance, or in the unicast network error correction instance. Conversely, we show that an arbitrary unicast secure network coding instance in which at most one link is eavesdropped can be reduced back to a multiple-unicast network coding instance. Additionally, we show that the capacity of a unicast network error correction instance in general is not (exactly) achievable. We derive upper bounds on the secrecy capacity for the secure network coding problem, based on cut-sets and the connectivity of links. Finally, we study optimal coding schemes for the network error correction problem, in the setting that the network and adversary parameters are not known a priori.
", "doi": "10.7907/Z9P26W5C", "publication_date": "2017", "thesis_type": "phd", "thesis_year": "2017" }, { "id": "thesis:10299", "collection": "thesis", "collection_id": "10299", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:06062017-101123885", "primary_object_url": { "basename": "Noorzad_Parham_2017_Thesis.pdf", "content": "final", "filesize": 1463018, "license": "other", "mime_type": "application/pdf", "url": "/10299/1/Noorzad_Parham_2017_Thesis.pdf", "version": "v4.0.0" }, "type": "thesis", "title": "Network Effects in Small Networks: A Study of Cooperation", "author": [ { "family_name": "Noorzad", "given_name": "Parham", "orcid": "0000-0002-0201-3791", "clpid": "Noorzad-Parham" } ], "thesis_advisor": [ { "family_name": "Effros", "given_name": "Michelle", "clpid": "Effros-M" } ], "thesis_committee": [ { "family_name": "Effros", "given_name": "Michelle", "clpid": "Effros-M" }, { "family_name": "Langberg", "given_name": "Michael", "clpid": "Langberg-M" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Hassibi", "given_name": "Babak", "clpid": "Hassibi-B" }, { "family_name": "Kostina", "given_name": "Victoria", "clpid": "Kostina-V" }, { "family_name": "Ligett", "given_name": "Katrina A.", "clpid": "Ligett-K-A" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "Communication over a point-to-point link is relatively well understood. However, when such a link is part of a larger network, our understanding is far from complete. Nonetheless, progress in this area has important consequences in both the theoretical and practical aspects of communication networks.
\r\n\r\nIn this work, we focus on the role of a single link in networks that in addition to point-to-point links, contain \"multi-terminal components.\" An example of a network consisting of a single multi-terminal component is the uplink in a wireless communication network where multiple transmitters communicate with a single receiver over a shared medium. We demonstrate the existence of a class of such networks where a finite capacity link results in a rate gain for each source that far exceeds the capacity of that link. This is an example of a \"network effect\": the phenomenon where a resource, here link capacity, is significantly more valuable in a network than in isolation. Here we measure the \"value\" of the finite capacity link by the sum-capacity gain per source that it enables.
\r\n\r\nThe central idea behind the construction of networks that exhibit such effects is the introduction of a node, referred to as the \"cooperation facilitator\" (CF), that allows other network nodes to work together to reduce interference. In the setting of the classical multiple access channel (MAC), an example of a CF is a node that receives rate-limited information from each transmitter and broadcasts rate-limited information back to the transmitters through a common bottleneck link. Let the \"cooperation rate\" be the capacity of the CF bottleneck link. We show that for a class of MACs, the presence of a CF\r\nleads to a sum-capacity gain that, as a function of the cooperation rate, has an infinite slope at cooperation rate zero. This means that the bottleneck link of the CF is significantly more valuable in some networks than in isolation. This class of MACs includes well-known examples such as the Gaussian MAC and the binary adder MAC.
\r\n\r\nIn addition to sum-capacity gain, cooperation under the CF model also improves reliability. Specifically, in the case of the MAC with two transmitters, whenever the CF has full access to both messages, the maximal- and average error capacity regions coincide. This effect is observed even when the cooperation rate is \"negligible\"; that is, the cooperation rate grows sublinearly in the number of channel uses. An implication of this result is the existence of a network whose maximal-error sum-capacity is not continuous with respect to the capacities of its edges; this means that in some networks, even a negligible cooperation rate leads to a positive sum-capacity gain.
", "doi": "10.7907/Z9M32STV", "publication_date": "2017", "thesis_type": "phd", "thesis_year": "2017" }, { "id": "thesis:10311", "collection": "thesis", "collection_id": "10311", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:06072017-154010713", "type": "thesis", "title": "A Study of Communication Networks through the Lens of Reduction", "author": [ { "family_name": "Wong", "given_name": "Ming Fai", "orcid": "0000-0002-9191-1277", "clpid": "Wong-Ming-Fai" } ], "thesis_advisor": [ { "family_name": "Effros", "given_name": "Michelle", "clpid": "Effros-M" } ], "thesis_committee": [ { "family_name": "Effros", "given_name": "Michelle", "clpid": "Effros-M" }, { "family_name": "Low", "given_name": "Steven H.", "clpid": "Low-S-H" }, { "family_name": "Langberg", "given_name": "Michael", "clpid": "Langberg-M" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Umans", "given_name": "Christopher M.", "clpid": "Umans-C-M" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "A central goal of information theory is to characterize the capacity regions of communication networks. Due to the difficulty of the general problem, research is primarily focused on families of problems defined by various classifiers. These classifiers include the channel transition function (i.e., noisy, deterministic, network coding), demand type (i.e., single-source, 2-unicast), network topology (i.e. acyclic network coding, index coding). To date, the families of networks that are fully solved remain limited. Moreover, results derived for one specific family often do not extend easily to other families of problems.
\r\n\r\nOur work shifts from the traditional focus on solving example networks to one that builds connections between problem solutions so that we can say where and when solving a problem in one domain would also solve a corresponding problem in another domain. Central to our approach is a technique called \"reduction\", in which we connect the solutions and results of communication problems. We say that problem A reduces to problem B when A can be solved by first transforming it to B and then applying a solution for B. We focus on two notions of reduction: reduction in code design and reduction in capacity region.
\r\n\r\nOur central results demonstrate reductions with respect to a variety of classifiers. We show that finding multiple multicast network capacity regions reduces to finding multiple unicast network capacity regions both when capacity is defined as the maximal rate over all possible codes and when capacity is defined as the optimal rate over linear codes. As a corollary to this result, we show that the same capacity reduction holds for when network types are limited to either network coding networks or index coding networks. In several instances, we show that a reduction in code design extends to a reduction in capacity region if and only if the edge removal conjecture holds. Here, the edge removal conjecture states that removing an edge of negligible capacity from a network does not change its capacity region.
\r\n\r\nOne of the key challenges in network coding research is how to handle networks containing cycles. As a result, many papers on network coding restrict attention to acyclic networks and some results derived for acyclic networks do not extend to networks containing cycles. We consider a streaming model for network communication where information is streamed to its destination under a constraint on maximal delay at the decoder. Restricting our attention to this scenario enables us to prove a code reduction from network coding to index coding in both acyclic and cyclic networks. Since index coding networks are acyclic, a consequence of this reduction is that under the streaming model, there is no fundamental difference between acyclic and cyclic networks.
", "doi": "10.7907/Z9348HFK", "publication_date": "2017", "thesis_type": "phd", "thesis_year": "2017" }, { "id": "thesis:10328", "collection": "thesis", "collection_id": "10328", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:06092017-013147699", "primary_object_url": { "basename": "Wael_Halbawi_phd_thesis.pdf", "content": "final", "filesize": 971642, "license": "other", "mime_type": "application/pdf", "url": "/10328/1/Wael_Halbawi_phd_thesis.pdf", "version": "v3.0.0" }, "type": "thesis", "title": "Error-Correcting Codes for Networks, Storage and Computation", "author": [ { "family_name": "Halbawi", "given_name": "Wael", "orcid": "0000-0001-5951-7002", "clpid": "Halbawi-Wael" } ], "thesis_advisor": [ { "family_name": "Hassibi", "given_name": "Babak", "clpid": "Hassibi-B" } ], "thesis_committee": [ { "family_name": "Hassibi", "given_name": "Babak", "clpid": "Hassibi-B" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Divsalar", "given_name": "Dariush", "clpid": "Divsalar-D" }, { "family_name": "Dimakis", "given_name": "Alexandros G.", "clpid": "Dimakis-A-G" }, { "family_name": "Kostina", "given_name": "Victoria", "clpid": "Kostina-V" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "The advent of the information age has bestowed upon us three challenges related to the way we deal with data. Firstly, there is an unprecedented demand for transmitting data at high rates. Secondly, the massive amounts of data being collected from various sources needs to be stored across time. Thirdly, there is a need to process the data collected and perform computations on it in order to extract meaningful information out of it. The interconnected nature of modern systems designed to perform these tasks has unraveled new difficulties when it comes to ensuring their resilience against sources of performance degradation. In the context of network communication and distributed data storage, system-level noise and adversarial errors have to be combated with efficient error correction schemes. In the case of distributed computation, the heterogeneous nature of computing clusters can potentially diminish the speedups promised by parallel algorithms, calling for schemes that mitigate the effect of slow machines and communication delay.
\r\n\r\nThis thesis addresses the problem of designing efficient fault tolerance schemes for the three scenarios just described. In the network communication setting, a family of multiple-source multicast networks that employ linear network coding is considered for which capacity-achieving distributed error-correcting codes, based on classical algebraic constructions, are designed. The codes require no coordination between the source nodes and are end to end: except for the source nodes and the destination node, the operation of the network remains unchanged.
\r\n\r\nIn the context of data storage, balanced error-correcting codes are constructed so that the encoding effort required is balanced out across the storage nodes. In particular, it is shown that for a fixed row weight, any cyclic Reed-Solomon code possesses a generator matrix in which the number of nonzeros is the same across the columns. In the balanced and sparsest case, where each row of the generator matrix is a minimum distance codeword, the maximal encoding time over the storage nodes is minimized, a property that is appealing in write-intensive settings. Analogous constructions are presented for a locally recoverable code construction due to Tamo and Barg.
\r\n\r\nLastly, the problem of mitigating stragglers in a distributed computation setup is addressed, where a function of some dataset is computed in parallel. Using Reed-Solomon coding techniques, a scheme is proposed that allows for the recovery of the function under consideration from the minimum number of machines possible. The only assumption made on the function is that it is additively separable, which renders the scheme useful in distributed gradient descent implementations. Furthermore, a theoretical model for the run time of the scheme is presented. When the return time of the machines is modeled probabilistically, the model can be used to optimally pick the scheme's parameters so that the expected computation time is minimized. The recovery is performed using an algorithm that runs in quadratic time and linear space, a notable improvement compared to state-of-the-art schemes.
\r\n\r\nThe unifying theme of the three scenarios is the construction of error-correcting codes whose encoding functions adhere to certain constraints. It is shown that in many cases, these constraints can be satisfied by classical constructions. As a result, the schemes presented are deterministic, operate over small finite fields and can be decoded using efficient algorithms.
", "doi": "10.7907/Z9J67F08", "publication_date": "2017", "thesis_type": "phd", "thesis_year": "2017" }, { "id": "thesis:9141", "collection": "thesis", "collection_id": "9141", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:09042015-171723764", "primary_object_url": { "basename": "MatthewThill2016_thesis.pdf", "content": "final", "filesize": 1775919, "license": "other", "mime_type": "application/pdf", "url": "/9141/1/MatthewThill2016_thesis.pdf", "version": "v4.0.0" }, "type": "thesis", "title": "Algebraic Techniques in Coding Theory: Entropy Vectors, Frames, and Constrained Coding", "author": [ { "family_name": "Thill", "given_name": "Matthew David", "orcid": "0000-0003-0885-6260", "clpid": "Thill-Matthew-David" } ], "thesis_advisor": [ { "family_name": "Hassibi", "given_name": "Babak", "clpid": "Hassibi-B" } ], "thesis_committee": [ { "family_name": "Hassibi", "given_name": "Babak", "clpid": "Hassibi-B" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Chandrasekaran", "given_name": "Venkat", "clpid": "Chandrasekaran-V" }, { "family_name": "Divsalar", "given_name": "Dariush", "clpid": "Divsalar-D" }, { "family_name": "Kostina", "given_name": "Victoria", "clpid": "Kostina-V" }, { "family_name": "Vaidyanathan", "given_name": "P. P.", "clpid": "Vaidyanathan-P-P" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "The study of codes, classically motivated by the need to communicate information reliably in the presence of error, has found new life in fields as diverse as network communication, distributed storage of data, and even has connections to the design of linear measurements used in compressive sensing. But in all contexts, a code typically involves exploiting the algebraic or geometric structure underlying an application. In this thesis, we examine several problems in coding theory, and try to gain some insight into the algebraic structure behind them.
\r\n\r\nThe first is the study of the entropy region - the space of all possible vectors of joint entropies which can arise from a set of discrete random variables. Understanding this region is essentially the key to optimizing network codes for a given network. To this end, we employ a group-theoretic method of constructing random variables producing so-called \"group-characterizable\" entropy vectors, which are capable of approximating any point in the entropy region. We show how small groups can be used to produce entropy vectors which violate the Ingleton inequality, a fundamental bound on entropy vectors arising from the random variables involved in linear network codes. We discuss the suitability of these groups to design codes for networks which could potentially outperform linear coding.
\r\n\r\nThe second topic we discuss is the design of frames with low coherence, closely related to finding spherical codes in which the codewords are unit vectors spaced out around the unit sphere so as to minimize the magnitudes of their mutual inner products. We show how to build frames by selecting a cleverly chosen set of representations of a finite group to produce a \"group code\" as described by Slepian decades ago. We go on to reinterpret our method as selecting a subset of rows of a group Fourier matrix, allowing us to study and bound our frames' coherences using character theory. We discuss the usefulness of our frames in sparse signal recovery using linear measurements.
\r\n\r\nThe final problem we investigate is that of coding with constraints, most recently motivated by the demand for ways to encode large amounts of data using error-correcting codes so that any small loss can be recovered from a small set of surviving data. Most often, this involves using a systematic linear error-correcting code in which each parity symbol is constrained to be a function of some subset of the message symbols. We derive bounds on the minimum distance of such a code based on its constraints, and characterize when these bounds can be achieved using subcodes of Reed-Solomon codes.
", "doi": "10.7907/Z9F18WNW", "publication_date": "2016", "thesis_type": "phd", "thesis_year": "2016" }, { "id": "thesis:8834", "collection": "thesis", "collection_id": "8834", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:04272015-133555770", "primary_object_url": { "basename": "Wei_Mao_Thesis_2015.pdf", "content": "final", "filesize": 1159546, "license": "other", "mime_type": "application/pdf", "url": "/8834/1/Wei_Mao_Thesis_2015.pdf", "version": "v2.0.0" }, "type": "thesis", "title": "Information-Theoretic Studies and Capacity Bounds: Group Network Codes and Energy Harvesting Communication Systems", "author": [ { "family_name": "Mao", "given_name": "Wei", "clpid": "Mao-Wei" } ], "thesis_advisor": [ { "family_name": "Hassibi", "given_name": "Babak", "clpid": "Hassibi-B" } ], "thesis_committee": [ { "family_name": "Hassibi", "given_name": "Babak", "clpid": "Hassibi-B" }, { "family_name": "Kostina", "given_name": "Victoria", "clpid": "Kostina-V" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Vaidyanathan", "given_name": "P. P.", "clpid": "Vaidyanathan-P-P" }, { "family_name": "Wierman", "given_name": "Adam C.", "clpid": "Wierman-A-C" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "Network information theory and channels with memory are two important but difficult frontiers of information theory. In this two-parted dissertation, we study these two areas, each comprising one part. For the first area we study the so-called entropy vectors via finite group theory, and the network codes constructed from finite groups. In particular, we identify the smallest finite group that violates the Ingleton inequality, an inequality respected by all linear network codes, but not satisfied by all entropy vectors. Based on the analysis of this group we generalize it to several families of Ingleton-violating groups, which may be used to design good network codes. Regarding that aspect, we study the network codes constructed with finite groups, and especially show that linear network codes are embedded in the group network codes constructed with these Ingleton-violating families. Furthermore, such codes are strictly more powerful than linear network codes, as they are able to violate the Ingleton inequality while linear network codes cannot. For the second area, we study the impact of memory to the channel capacity through a novel communication system: the energy harvesting channel. Different from traditional communication systems, the transmitter of an energy harvesting channel is powered by an exogenous energy harvesting device and a finite-sized battery. As a consequence, each time the system can only transmit a symbol whose energy consumption is no more than the energy currently available. This new type of power supply introduces an unprecedented input constraint for the channel, which is random, instantaneous, and has memory. Furthermore, naturally, the energy harvesting process is observed causally at the transmitter, but no such information is provided to the receiver. Both of these features pose great challenges for the analysis of the channel capacity. In this work we use techniques from channels with side information, and finite state channels, to obtain lower and upper bounds of the energy harvesting channel. In particular, we study the stationarity and ergodicity conditions of a surrogate channel to compute and optimize the achievable rates for the original channel. In addition, for practical code design of the system we study the pairwise error probabilities of the input sequences.", "doi": "10.7907/Z9ZS2TFB", "publication_date": "2015", "thesis_type": "phd", "thesis_year": "2015" }, { "id": "thesis:8814", "collection": "thesis", "collection_id": "8814", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:04082015-142940694", "type": "thesis", "title": "Rewriting Schemes for Flash Memory", "author": [ { "family_name": "En Gad", "given_name": "Eyal", "clpid": "En-Gad-Eyal" } ], "thesis_advisor": [ { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" } ], "thesis_committee": [ { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Effros", "given_name": "Michelle", "clpid": "Effros-M" }, { "family_name": "Vaidyanathan", "given_name": "P. P.", "clpid": "Vaidyanathan-P-P" }, { "family_name": "Hassibi", "given_name": "Babak", "clpid": "Hassibi-B" }, { "family_name": "Langberg", "given_name": "Michael", "clpid": "Langberg-M" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "Flash memory is a leading storage media with excellent features such as random access and\r\nhigh storage density. However, it also faces significant reliability and endurance challenges.\r\nIn flash memory, the charge level in the cells can be easily increased, but removing charge\r\nrequires an expensive erasure operation. In this thesis we study rewriting schemes that\r\nenable the data stored in a set of cells to be rewritten by only increasing the charge level\r\nin the cells. We consider two types of modulation scheme; a convectional modulation based\r\non the absolute levels of the cells, and a recently-proposed scheme based on the relative cell\r\nlevels, called rank modulation. The contributions of this thesis to the study of rewriting\r\nschemes for rank modulation include the following: we
\r\n\r\n•propose a new method of rewriting in rank modulation, beyond the previously proposed\r\nmethod of \u201cpush-to-the-top\u201d;
\r\n•study the limits of rewriting with the newly proposed method, and derive a tight upper\r\nbound of 1 bit per cell;
\r\n•extend the rank-modulation scheme to support rankings with repetitions, in order to\r\nimprove the storage density;
\r\n•derive a tight upper bound of 2 bits per cell for rewriting in rank modulation with\r\nrepetitions;
\r\n•construct an efficient rewriting scheme that asymptotically approaches the upper bound\r\nof 2 bit per cell.
\r\n\r\nThe next part of this thesis studies rewriting schemes for a conventional absolute-levels\r\nmodulation. The considered model is called \u201cwrite-once memory\u201d (WOM). We focus on\r\nWOM schemes that achieve the capacity of the model. In recent years several capacity-achieving\r\nWOM schemes were proposed, based on polar codes and randomness extractors.\r\nThe contributions of this thesis to the study of WOM scheme include the following: we
\r\n\r\n•propose a new capacity-achievingWOM scheme based on sparse-graph codes, and show\r\nits attractive properties for practical implementation;
\r\n•improve the design of polarWOMschemes to remove the reliance on shared randomness\r\nand include an error-correction capability.
\r\n\r\nThe last part of the thesis studies the local rank-modulation (LRM) scheme, in which a\r\nsliding window going over a sequence of real-valued variables induces a sequence of permutations.\r\nThe LRM scheme is used to simulate a single conventional multi-level flash cell.\r\nThe simulated cell is realized by a Gray code traversing all the relative-value states where,\r\nphysically, the transition between two adjacent states in the Gray code is achieved by using\r\na single \u201cpush-to-the-top\u201d operation. The main results of the last part of the thesis are two\r\nconstructions of Gray codes with asymptotically-optimal rate.
", "doi": "10.7907/Z9R49NQ3", "publication_date": "2015", "thesis_type": "phd", "thesis_year": "2015" }, { "id": "thesis:8888", "collection": "thesis", "collection_id": "8888", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:05262015-094933189", "primary_object_url": { "basename": "CarlosGonzalez2015_thesis.pdf", "content": "final", "filesize": 1591603, "license": "other", "mime_type": "application/pdf", "url": "/8888/73/CarlosGonzalez2015_thesis.pdf", "version": "v9.0.0" }, "type": "thesis", "title": "Optimal Data Distributions in Machine Learning", "author": [ { "family_name": "Gonz\u00e1lez Palacios", "given_name": "Carlos Roberto", "clpid": "Gonz\u00e1lez-Palacios-Carlos-Roberto" } ], "thesis_advisor": [ { "family_name": "Abu-Mostafa", "given_name": "Yaser S.", "clpid": "Abu-Mostafa-Y-S" }, { "family_name": "Perona", "given_name": "Pietro", "clpid": "Perona-P" } ], "thesis_committee": [ { "family_name": "Abu-Mostafa", "given_name": "Yaser S.", "clpid": "Abu-Mostafa-Y-S" }, { "family_name": "Perona", "given_name": "Pietro", "clpid": "Perona-P" }, { "family_name": "Vaidyanathan", "given_name": "P. P.", "clpid": "Vaidyanathan-P-P" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Yue", "given_name": "Yisong", "clpid": "Yue-Yisong" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "In the first part of the thesis we explore three fundamental questions that arise naturally when we conceive a machine learning scenario where the training and test distributions can differ. Contrary to conventional wisdom, we show that in fact mismatched training and test distribution can yield better out-of-sample performance. This optimal performance can be obtained by training with the dual distribution. This optimal training distribution depends on the test distribution set by the problem, but not on the target function that we want to learn. We show how to obtain this distribution in both discrete and continuous input spaces, as well as how to approximate it in a practical scenario. Benefits of using this distribution are exemplified in both synthetic and real data sets.
\r\n\r\nIn order to apply the dual distribution in the supervised learning scenario where the training data set is fixed, it is necessary to use weights to make the sample appear as if it came from the dual distribution. We explore the negative effect that weighting a sample can have. The theoretical decomposition of the use of weights regarding its effect on the out-of-sample error is easy to understand but not actionable in practice, as the quantities involved cannot be computed. Hence, we propose the Targeted Weighting algorithm that determines if, for a given set of weights, the out-of-sample performance will improve or not in a practical setting. This is necessary as the setting assumes there are no labeled points distributed according to the test distribution, only unlabeled samples.
\r\n\r\nFinally, we propose a new class of matching algorithms that can be used to match the training set to a desired distribution, such as the dual distribution (or the test distribution). These algorithms can be applied to very large datasets, and we show how they lead to improved performance in a large real dataset such as the Netflix dataset. Their computational complexity is the main reason for their advantage over previous algorithms proposed in the covariate shift literature.
\r\n\r\nIn the second part of the thesis we apply Machine Learning to the problem of behavior recognition. We develop a specific behavior classifier to study fly aggression, and we develop a system that allows analyzing behavior in videos of animals, with minimal supervision. The system, which we call CUBA (Caltech Unsupervised Behavior Analysis), allows detecting movemes, actions, and stories from time series describing the position of animals in videos. The method summarizes the data, as well as it provides biologists with a mathematical tool to test new hypotheses. Other benefits of CUBA include finding classifiers for specific behaviors without the need for annotation, as well as providing means to discriminate groups of animals, for example, according to their genetic line.
", "doi": "10.7907/Z9DR2SD5", "publication_date": "2015", "thesis_type": "phd", "thesis_year": "2015" }, { "id": "thesis:8041", "collection": "thesis", "collection_id": "8041", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:12132013-161604752", "type": "thesis", "title": "Quantifying Synergistic Information", "author": [ { "family_name": "Griffith", "given_name": "Virgil", "clpid": "Griffith-Virgil" } ], "thesis_advisor": [ { "family_name": "Koch", "given_name": "Christof", "orcid": "0000-0001-6482-8067", "clpid": "Koch-C" } ], "thesis_committee": [ { "family_name": "Perona", "given_name": "Pietro", "orcid": "0000-0002-7583-5809", "clpid": "Perona-P" }, { "family_name": "Koch", "given_name": "Christof", "orcid": "0000-0001-6482-8067", "clpid": "Koch-C" }, { "family_name": "Ho", "given_name": "Tracey C.", "clpid": "Ho-Tracey" }, { "family_name": "Beck", "given_name": "James L.", "clpid": "Beck-J-L" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" } ], "local_group": [ { "literal": "Koch Laboratory (KLAB)" }, { "literal": "div_eng" } ], "abstract": "Within the microcosm of information theory, I explore what it means for a system to be functionally irreducible. This is operationalized as quantifying the extent to which cooperative or \"synergistic\" effects enable random variables X1, ... , Xn to predict (have mutual information about) a single target random variable Y . In Chapter 1, we introduce the problem with some emblematic examples. In Chapter 2, we show how six different measures from the existing literature fail to quantify this notion of synergistic mutual information. In Chapter 3 we take a step towards a measure of synergy which yields the first nontrivial lowerbound on synergistic mutual information. In Chapter 4, we find that synergy is but the weakest notion of a broader concept of irreducibility. In Chapter 5, we apply our results from Chapters 3 and 4 towards grounding Giulio Tononi\u2019s ambitious \u03c6 measure which attempts to quantify the magnitude of consciousness experience.
", "doi": "10.7907/ZS2T-XQ55", "publication_date": "2014", "thesis_type": "phd", "thesis_year": "2014" }, { "id": "thesis:8041", "collection": "thesis", "collection_id": "8041", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:12132013-161604752", "type": "thesis", "title": "Quantifying Synergistic Information", "author": [ { "family_name": "Griffith", "given_name": "Virgil", "clpid": "Griffith-Virgil" } ], "thesis_advisor": [ { "family_name": "Koch", "given_name": "Christof", "orcid": "0000-0001-6482-8067", "clpid": "Koch-C" } ], "thesis_committee": [ { "family_name": "Perona", "given_name": "Pietro", "orcid": "0000-0002-7583-5809", "clpid": "Perona-P" }, { "family_name": "Koch", "given_name": "Christof", "orcid": "0000-0001-6482-8067", "clpid": "Koch-C" }, { "family_name": "Ho", "given_name": "Tracey C.", "clpid": "Ho-Tracey" }, { "family_name": "Beck", "given_name": "James L.", "clpid": "Beck-J-L" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" } ], "local_group": [ { "literal": "Koch Laboratory (KLAB)" }, { "literal": "div_eng" } ], "abstract": "Within the microcosm of information theory, I explore what it means for a system to be functionally irreducible. This is operationalized as quantifying the extent to which cooperative or \"synergistic\" effects enable random variables X1, ... , Xn to predict (have mutual information about) a single target random variable Y . In Chapter 1, we introduce the problem with some emblematic examples. In Chapter 2, we show how six different measures from the existing literature fail to quantify this notion of synergistic mutual information. In Chapter 3 we take a step towards a measure of synergy which yields the first nontrivial lowerbound on synergistic mutual information. In Chapter 4, we find that synergy is but the weakest notion of a broader concept of irreducibility. In Chapter 5, we apply our results from Chapters 3 and 4 towards grounding Giulio Tononi\u2019s ambitious \u03c6 measure which attempts to quantify the magnitude of consciousness experience.
", "doi": "10.7907/ZS2T-XQ55", "publication_date": "2014", "thesis_type": "phd", "thesis_year": "2014" }, { "id": "thesis:8232", "collection": "thesis", "collection_id": "8232", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:05132014-142306756", "type": "thesis", "title": "Crystals that Count! Physical Principles and Experimental Investigations of DNA Tile Self-Assembly", "author": [ { "family_name": "Evans", "given_name": "Constantine Glen", "orcid": "0000-0002-7053-1670", "clpid": "Evans-Constantine-Glen" } ], "thesis_advisor": [ { "family_name": "Winfree", "given_name": "Erik", "clpid": "Winfree-E" } ], "thesis_committee": [ { "family_name": "Winfree", "given_name": "Erik", "clpid": "Winfree-E" }, { "family_name": "Libbrecht", "given_name": "Kenneth George", "clpid": "Libbrecht-K-G" }, { "family_name": "Rothemund", "given_name": "Paul W. K.", "clpid": "Rothemund-P-W-K" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Pierce", "given_name": "Niles A.", "clpid": "Pierce-N-A" } ], "local_group": [ { "literal": "div_pma" } ], "abstract": "Algorithmic DNA tiles systems are fascinating. From a theoretical perspective, they can result in simple systems that assemble themselves into beautiful, complex structures through fundamental interactions and logical rules. As an experimental technique, they provide a promising method for programmably assembling complex, precise crystals that can grow to considerable size while retaining nanoscale resolution. In the journey from theoretical abstractions to experimental demonstrations, however, lie numerous challenges and complications.
\r\n\r\nIn this thesis, to examine these challenges, we consider the physical principles behind DNA tile self-assembly. We survey recent progress in experimental algorithmic self-assembly, and explain the simple physical models behind this progress. Using direct observation of individual tile attachments and detachments with an atomic force microscope, we test some of the fundamental assumptions of the widely-used kinetic Tile Assembly Model, obtaining results that fit the model to within error. We then depart from the simplest form of that model, examining the effects of DNA sticky end sequence energetics on tile system behavior. We develop theoretical models, sequence assignment algorithms, and a software package, StickyDesign, for sticky end sequence design.
\r\n\r\nAs a demonstration of a specific tile system, we design a binary counting ribbon that can accurately count from a programmable starting value and stop growing after overflowing, resulting in a single system that can construct ribbons of precise and programmable length. In the process of designing the system, we explain numerous considerations that provide insight into more general tile system design, particularly with regards to tile concentrations, facet nucleation, the construction of finite assemblies, and design beyond the abstract Tile Assembly Model.
\r\n\r\nFinally, we present our crystals that count: experimental results with our binary counting system that represent a significant improvement in the accuracy of experimental algorithmic self-assembly, including crystals that count perfectly with 5 bits from 0 to 31. We show some preliminary experimental results on the construction of our capping system to stop growth after counters overflow, and offer some speculation on potential future directions of the field.
\r\n", "doi": "10.7907/7FMK-9403", "publication_date": "2014", "thesis_type": "phd", "thesis_year": "2014" }, { "id": "thesis:7460", "collection": "thesis", "collection_id": "7460", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:02042013-110332492", "primary_object_url": { "basename": "JosephSchaeffer_PhD_Thesis_Complete.pdf", "content": "final", "filesize": 4045542, "license": "other", "mime_type": "application/pdf", "url": "/7460/1/JosephSchaeffer_PhD_Thesis_Complete.pdf", "version": "v2.0.0" }, "type": "thesis", "title": "Stochastic Simulation of the Kinetics of Multiple Interacting Nucleic Acid Strands", "author": [ { "family_name": "Schaeffer", "given_name": "Joseph Malcolm", "clpid": "Schaeffer-Joseph-Malcolm" } ], "thesis_advisor": [ { "family_name": "Winfree", "given_name": "Erik", "clpid": "Winfree-E" } ], "thesis_committee": [ { "family_name": "Winfree", "given_name": "Erik", "clpid": "Winfree-E" }, { "family_name": "Pierce", "given_name": "Niles A.", "clpid": "Pierce-N-A" }, { "family_name": "Umans", "given_name": "Christopher M.", "clpid": "Umans-C-M" }, { "family_name": "Barr", "given_name": "Alan H.", "clpid": "Barr-A-H" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "DNA nanotechnology is an emerging field which utilizes the unique structural properties of nucleic acids in order to build nanoscale devices, such as logic gates, motors, walkers, and algorithmic structures. Predicting the structure and interactions of a DNA device requires good modeling of both the thermodynamics and the kinetics of the DNA strands within the system. The kinetics of a set of DNA strands can be modeled as a continuous time Markov process through the state space of all secondary structures. The primary means of exploring the kinetics of a DNA system is by simulating trajectories through the state space and aggregating data over many such trajectories.
\r\n\r\nWe expand on previous work by extending the thermodynamics and kinetics models to handle multiple strands in a fixed volume, and show that the new models are consistent with previous models. We developed data structures and algorithms that allow us to take advantage of local properties of secondary structure, improving the efficiency of the simulator so that we can handle larger systems. The new kinetic parameters in our model were calibrated by analyzing simulator results on experimental systems that measure basic kinetic rates of various processes. Finally, we apply the new simulator to explore a case study on toehold-mediated four-way branch migration.
", "doi": "10.7907/JEBY-6X69", "publication_date": "2013", "thesis_type": "phd", "thesis_year": "2013" }, { "id": "thesis:7176", "collection": "thesis", "collection_id": "7176", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:07122012-141803264", "primary_object_url": { "basename": "Zhou_Hongchao_2013_Thesis.pdf", "content": "final", "filesize": 3182883, "license": "other", "mime_type": "application/pdf", "url": "/7176/1/Zhou_Hongchao_2013_Thesis.pdf", "version": "v4.0.0" }, "type": "thesis", "title": "Randomness and Noise in Information Systems", "author": [ { "family_name": "Zhou", "given_name": "Hongchao", "clpid": "Zhou-Hongchao" } ], "thesis_advisor": [ { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" } ], "thesis_committee": [ { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Winfree", "given_name": "Erik", "clpid": "Winfree-E" }, { "family_name": "Hassibi", "given_name": "Babak", "clpid": "Hassibi-B" }, { "family_name": "Umans", "given_name": "Christopher M.", "clpid": "Umans-C-M" }, { "family_name": "Effros", "given_name": "Michelle", "clpid": "Effros-M" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "This dissertation is devoted to the study of randomness and noise in a number of information systems including computation systems, storage systems, and natural paradigms like molecular systems, where randomness plays important and distinct roles. Motivated by applications in engineering and science we address a number of theoretical research questions.\r\nStorage systems are widely used and have played a crucial rule in both consumer and industrial products, for example, personal computers, data centers, and embedded systems. However, such system suffers from issues of cost, restricted-lifetime, and reliability with the emergence of new systems and devices, such as distributed storage and flash memory, respectively. Information theory, on the other hand, provides fundamental bounds and solutions to fully utilize resources such as data density, information I/O and network bandwidth. This thesis bridges these two topics, and proposes to solve challenges in data storage using a variety of coding techniques, so that storage becomes faster, more affordable, and more reliable.
\r\n\r\nWe consider the system level and study the integration of RAID schemes and distributed storage. Erasure-correcting codes are the basis of the ubiquitous RAID schemes for storage systems, where disks correspond to symbols in the code and are located in a (distributed) network. Specifically, RAID schemes are based on MDS (maximum distance separable) array codes that enable optimal storage and efficient encoding and decoding algorithms. With r redundancy symbols an MDS code can sustain r erasures. For example, consider an MDS code that can correct two erasures. It is clear that when two symbols are erased, one needs to access and transmit all the remaining information to rebuild the erasures. However, an interesting and practical question is: What is the smallest fraction of information that one needs to access and transmit in order to correct a single erasure? In Part I we will show that the lower bound of 1/2 is achievable and that the result can be generalized to codes with arbitrary number of parities and optimal rebuilding.
\r\n\r\nWe consider the device level and study coding and modulation techniques for emerging non-volatile memories such as flash memory. In particular, rank modulation is a novel data representation scheme proposed by Jiang et al. for multi-level flash memory cells, in which a set of n cells stores information in the permutation induced by the different charge levels of the individual cells. It eliminates the need for discrete cell levels, as well as overshoot errors, when programming cells. In order to decrease the decoding complexity, we propose two variations of this scheme in Part II: bounded rank modulation where only small sliding windows of cells are sorted to generated permutations, and partial rank modulation where only part of the n cells are used to represent data. We study limits on the capacity of bounded rank modulation and propose encoding and decoding algorithms. We show that overlaps between windows will increase capacity. We present Gray codes spanning all possible partial-rank states and using only ``push-to-the-top'' operations. These Gray codes turn out to solve an open combinatorial problem called universal cycle, which is a sequence of integers generating all possible partial permutations.
\r\n", "doi": "10.7907/TFHZ-RW88", "publication_date": "2013", "thesis_type": "phd", "thesis_year": "2013" }, { "id": "thesis:7355", "collection": "thesis", "collection_id": "7355", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:12282012-075056374", "type": "thesis", "title": "On the Analysis and Design of the Locust Olfactory System", "author": [ { "family_name": "Tootoonian", "given_name": "Sina", "clpid": "Tootoonian-Sina" } ], "thesis_advisor": [ { "family_name": "Laurent", "given_name": "Gilles J.", "orcid": "0000-0002-2296-114X", "clpid": "Laurent-G-J" } ], "thesis_committee": [ { "family_name": "Siapas", "given_name": "Athanassios G.", "orcid": "0000-0001-8837-678X", "clpid": "Siapas-A-G" }, { "family_name": "Perona", "given_name": "Pietro", "orcid": "0000-0002-7583-5809", "clpid": "Perona-P" }, { "family_name": "Schuman", "given_name": "Erin Margaret", "orcid": "0000-0002-7053-1005", "clpid": "Schuman-E-M" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Laurent", "given_name": "Gilles J.", "orcid": "0000-0002-2296-114X", "clpid": "Laurent-G-J" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "The ~ 830 projection neurons (PNs) of the locust antennal lobe respond to odors with dense, odor-specific spatio-temporal activity patterns that are mapped via intrinsic and circuit properties into a sparse representation by the Kenyon cells of the mushroom body, which are in turn read out by the beta-lobe neurons (bLNs). In this thesis we present several analyses of this system. First, we describe metrics for quantifying the geometric properties of PN population responses in the full response space that verify the structures revealed by locally linear embedding. Second, we analyze the mixture responses of single PNs and find that in many cases the mixture response can be explained using one of the component responses. Grouping PNs by their single component preferences reveals a potentially simple substrate for olfactory computations. Third, we look for evidence of cycle-by-cycle decoding of PNs by KCs. We show that much of the variance in single KC responses can be explained using small numbers of PNs, and conversely, that PN odor response trajectories can be reconstructed using KC responses. Finally, in a theoretical / computational analysis, we assemble some of the basic biological facts about the locust olfactory system into an architecture for the online learning of arbitrary mappings from odors to valences.", "doi": "10.7907/MAWA-GD04", "publication_date": "2013", "thesis_type": "phd", "thesis_year": "2013" }, { "id": "thesis:7148", "collection": "thesis", "collection_id": "7148", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:06082012-122324439", "type": "thesis", "title": "Network Coding and Distributed Compression over Large Networks: Some Basic Principles", "author": [ { "family_name": "Bakshi", "given_name": "Mayank", "clpid": "Bakshi-Mayank" } ], "thesis_advisor": [ { "family_name": "Effros", "given_name": "Michelle", "orcid": "0000-0003-3757-0675", "clpid": "Effros-M" } ], "thesis_committee": [ { "family_name": "Effros", "given_name": "Michelle", "orcid": "0000-0003-3757-0675", "clpid": "Effros-M" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Ho", "given_name": "Tracey C.", "clpid": "Ho-Tracey" }, { "family_name": "Chandy", "given_name": "K. Mani", "orcid": "0000-0001-9190-1290", "clpid": "Chandy-K-M" }, { "family_name": "Wierman", "given_name": "Adam C.", "orcid": "0000-0002-5923-0199", "clpid": "Wierman-A-C" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "The fields of Network Coding and Distributed Compression have focused primarily on finding the capacity for families of problems defined by either a broad class of networks topologies (e.g., directed, acyclic networks) under a narrow class of demands (e.g., multicast), or a specific network topology (e.g. three-node networks) under different types of demands (e.g. Slepian-Wolf, Ahlswede-K\u00f6rner). Given the difficulty of the general problem, it is not surprising that the collection of networks that have been fully solved to date is still very small. This work investigates several new approaches to bounding the achievable rate region for general network source coding problems - reducing a network to an equivalent network or collection of networks, investigating the effect of feedback on achievable rates, and characterizing the role of side information.
\r\n\r\nWe describe two approaches aimed at simplifying the capacity calculations in a large network. First, we prove the optimality of separation between network coding and channel coding for networks of point-to-point channels with a Byzantine adversary. Next, we give a strategy for calculating the capacity of an error-free network by decomposing that network into smaller networks. We show that this strategy is optimal for a large class of networks and give a bound for other cases.
\r\n\r\nTo date, the role of feedback in network source coding has received very little attention. We present several examples of networks that demonstrate that feedback can increases the set of achievable rates in both lossy and lossless network source coding settings. We derive general upper and lower bounds on the rate regions for networks with limited feedback that demonstrate a fundamental tradeoff between the forward rate and the feedback rate. For zero error source coding with limited feedback and decoder side information, we derive the exact tradeoff between the forward rate and the feedback rate for several classes of sources. A surprising result is that even zero rate feedback can reduce the optimal forward rate by an arbitrary factor.
\r\n\r\nSide information can be used to reduce the rates required for reliable information. We precisely characterize the exact achievable region for multicast networks with side information at the sinks and find upper and lower bounds on the achievable rate region for other demand types.
", "doi": "10.7907/GWDW-5H78", "publication_date": "2012", "thesis_type": "phd", "thesis_year": "2012" }, { "id": "thesis:7105", "collection": "thesis", "collection_id": "7105", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:05302012-200743814", "primary_object_url": { "basename": "JonathanHarel_PhD_Thesis.pdf", "content": "final", "filesize": 23919523, "license": "other", "mime_type": "application/pdf", "url": "/7105/1/JonathanHarel_PhD_Thesis.pdf", "version": "v8.0.0" }, "type": "thesis", "title": "Neural Pattern Similarity and Visual Perception", "author": [ { "family_name": "Harel", "given_name": "Jonathan", "clpid": "Harel-Jonathan" } ], "thesis_advisor": [ { "family_name": "Koch", "given_name": "Christof", "orcid": "0000-0001-6482-8067", "clpid": "Koch-C" }, { "family_name": "Adolphs", "given_name": "Ralph", "orcid": "0000-0002-8053-9692", "clpid": "Adolphs-R" } ], "thesis_committee": [ { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Koch", "given_name": "Christof", "orcid": "0000-0001-6482-8067", "clpid": "Koch-C" }, { "family_name": "Adolphs", "given_name": "Ralph", "orcid": "0000-0002-8053-9692", "clpid": "Adolphs-R" }, { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" }, { "family_name": "Perona", "given_name": "Pietro", "orcid": "0000-0002-7583-5809", "clpid": "Perona-P" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "This thesis addresses the question of whether people actually see the same visual stimuli somehow differently, and under what conditions, if so. It is an experimental contribution to the basic understanding of visual and especially face perception, and its neural correlates, with an emphasis on comparing patterns of neural activity driven by visual stimuli across trials and across individuals. We make extensive use of functional magnetic resonance imaging (fMRI); all inferences about neural activity are made via this intermediary. The thesis is organized into two parts:
\r\n\r\nIn Part I, we investigate the nature of face familiarity and distinctiveness at perceptual and neural levels. We first address the question of how the faces of those people personally familiar to a viewer appear different than they would to an unfamiliar viewer. The main result is that they appear more distinctive, i.e., dissimilar to and distinguishable from other faces, and more so the higher the level of familiarity. Having established this connection between face familiarity and distinctiveness, we ask next what is different about the perception of such faces, as compared with indistinct and unfamiliar faces, at the level of brain activation. We find that familiar and distinctive faces are represented more consistently: compared with indistinct faces, which evoke slightly different patterns of activity with each new presentation, these faces evoke slightly similar patterns. Combined with the observation that consistency can enhance memory encoding (a result reported by Xue et al. [102]), this suggests a cyclic process for the learning of unfamiliar faces in which consistent representation and the presence of newly formed memories mutually feedback on each other.
\r\n\r\nWhereas in Part I we focus on individual differences in neural activity, principally by experimentally manipulating stimulus familiarity, in Part II, we shift our focus to similarities across individuals and extend our investigation beyond faces to the perception of visual objects in general and moving images. We begin with an experiment involving the perception of static images selected from 44 object categories, where we find that the distances between these categories, induced from activity in cortical visual object areas, correlate highly between subjects, and also to distances inferred from a behavioral clustering task, and that this correlation remains significant even among subsets of closely related categories. We also show that one subject's brain activity can be accurately modeled using another's, and that this allows us to predict which image a subject is viewing based on his/her brain activity. Then, in a different experiment investigating the perception of dynamic/video stimuli, we find evidence that when watching videos with sound, visual attention is likely blurred at times and transferred to audition; subjects relatively temporally decorrelate in visual areas compared to the muted case, in which the patterns of neural activity correlate across subjects at an average of 78% the level found with oneself later in time.
\r\n\r\nThe findings reported in this thesis thus offer quantitative lower bounds on how similarly different individuals neurally experience visual stimuli, and an explanation for how they perceptually and neurally diverge when familiarity with a (face) stimulus varies, suggesting a possible mechanism for the encoding of new visual objects into memory. We conclude with a discussion of some of the questions raised by this work and directions for future research.
", "doi": "10.7907/1BCV-MQ11", "publication_date": "2012", "thesis_type": "phd", "thesis_year": "2012" }, { "id": "thesis:7064", "collection": "thesis", "collection_id": "7064", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:05222012-145639265", "primary_object_url": { "basename": "bodinebaron0522.pdf", "content": "final", "filesize": 6129584, "license": "other", "mime_type": "application/pdf", "url": "/7064/1/bodinebaron0522.pdf", "version": "v6.0.0" }, "type": "thesis", "title": "Peer Effects in Social Networks: Search, Matching Markets, and Epidemics", "author": [ { "family_name": "Bodine-Baron", "given_name": "Elizabeth Anne", "clpid": "Bodine-Baron-Elizabeth-Anne" } ], "thesis_advisor": [ { "family_name": "Hassibi", "given_name": "Babak", "clpid": "Hassibi-B" }, { "family_name": "Wierman", "given_name": "Adam C.", "clpid": "Wierman-A-C" } ], "thesis_committee": [ { "family_name": "Hassibi", "given_name": "Babak", "clpid": "Hassibi-B" }, { "family_name": "Low", "given_name": "Steven H.", "clpid": "Low-S-H" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Yariv", "given_name": "Leeat", "clpid": "Yariv-L" }, { "family_name": "Wierman", "given_name": "Adam C.", "clpid": "Wierman-A-C" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "Social network analysis emerged as an important area in sociology in the early 1930s, marking a shift from looking at individual attribute data to examining the relationships between people and groups. Surveying many different types of real-world networks, researchers quickly found that different types of social networks tend to share a common set of structural characteristics, including small diameter, high clustering, and heavy-tailed degree distributions. Moving beyond real networks, in the 1990s researchers began to propose random network models to explain these commonly observed social network structures. These models laid the foundation for investigation into problems where the underlying network plays a key role, from the spread of information and disease, to the design of distributed communication and search algorithms, to mechanism design and public policy. Here we focus on the role of peer effects in social networks. Through this lens, we develop a mathematically tractable random network model incorporating searchability, propose a novel way to model and analyze two-sided matching markets with externalities, model and calculate the cost of an epidemic spreading on a complex network, and examine the impact of conforming and non-conforming peer effects in vaccination decisions on public health policy.
\r\n\r\nThroughout this work, the goal is to bring together knowledge and techniques from diverse fields like sociology, engineering, and economics, exploiting our understanding of social network structure and generative models to understand deeper problems that \u2014 without this knowledge \u2014 could be intractable. Instead of crippling our analysis, social network characteristics allow us to reach deeper insights about the interaction between a particular problem and the network underlying it.
", "doi": "10.7907/GDV2-YF12", "publication_date": "2012", "thesis_type": "phd", "thesis_year": "2012" }, { "id": "thesis:6953", "collection": "thesis", "collection_id": "6953", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:04202012-133210844", "primary_object_url": { "basename": "rolfe_jason_2012_thesis.pdf", "content": "final", "filesize": 10201646, "license": "other", "mime_type": "application/pdf", "url": "/6953/1/rolfe_jason_2012_thesis.pdf", "version": "v4.0.0" }, "type": "thesis", "title": "Intrinsic Gradient Networks", "author": [ { "family_name": "Rolfe", "given_name": "Jason Tyler", "clpid": "Rolfe-Jason-Tyler" } ], "thesis_advisor": [ { "family_name": "Cook", "given_name": "Matthew M.", "clpid": "Cook-M-M" }, { "family_name": "Perona", "given_name": "Pietro", "orcid": "0000-0002-7583-5809", "clpid": "Perona-P" } ], "thesis_committee": [ { "family_name": "Perona", "given_name": "Pietro", "orcid": "0000-0002-7583-5809", "clpid": "Perona-P" }, { "family_name": "Cook", "given_name": "Matthew M.", "clpid": "Cook-M-M" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Winfree", "given_name": "Erik", "orcid": "0000-0002-5899-7523", "clpid": "Winfree-E" }, { "family_name": "Koch", "given_name": "Christof", "orcid": "0000-0001-6482-8067", "clpid": "Koch-C" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "Artificial neural networks are computationally powerful and exhibit brain-like dynamics. Unfortunately, the conventional gradient-dependent learning algorithms used to train them are biologically implausible. The calculation of the gradient in a traditional artificial neural network requires a complementary network of fast training signals that are dependent upon, but must not affect, the primary output-generating network activity. In contrast, the network of neurons in the cortex is highly recurrent; a network of gradient-calculating neurons in the brain would certainly project back to and influence the primary network. We address this biological implausibility by introducing a novel class of recurrent neural networks, intrinsic gradient networks, for which the gradient of an error function with respect to the parameters is a simple function of the network state. These networks can be trained using only their intrinsic signals, much like the network of neurons in the brain.
\r\n\r\nWe derive a simple equation that characterizes intrinsic gradient networks, and construct a broad set of networks that satisfy this characteristic equation. The resulting set of intrinsic gradient networks includes many highly recurrent instances for which the gradient can be calculated by a simple, local, pseudo-Hebbian function of the network state, thus resolving a long-standing contradiction between artificial and biological neural networks. We demonstrate that these networks can learn to perform nontrivial tasks like handwritten digit recognition using only their intrinsic signals. Finally, we show that a cortical implementation of an intrinsic gradient network would have a number of characteristic computational, anatomical, and electrophysiological properties, and review experimental evidence suggesting the manifestation of these properties in the cortex.
", "doi": "10.7907/YCB7-7X24", "publication_date": "2012", "thesis_type": "phd", "thesis_year": "2012" }, { "id": "thesis:6953", "collection": "thesis", "collection_id": "6953", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:04202012-133210844", "primary_object_url": { "basename": "rolfe_jason_2012_thesis.pdf", "content": "final", "filesize": 10201646, "license": "other", "mime_type": "application/pdf", "url": "/6953/1/rolfe_jason_2012_thesis.pdf", "version": "v4.0.0" }, "type": "thesis", "title": "Intrinsic Gradient Networks", "author": [ { "family_name": "Rolfe", "given_name": "Jason Tyler", "clpid": "Rolfe-Jason-Tyler" } ], "thesis_advisor": [ { "family_name": "Cook", "given_name": "Matthew M.", "clpid": "Cook-M-M" }, { "family_name": "Perona", "given_name": "Pietro", "orcid": "0000-0002-7583-5809", "clpid": "Perona-P" } ], "thesis_committee": [ { "family_name": "Perona", "given_name": "Pietro", "orcid": "0000-0002-7583-5809", "clpid": "Perona-P" }, { "family_name": "Cook", "given_name": "Matthew M.", "clpid": "Cook-M-M" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Winfree", "given_name": "Erik", "orcid": "0000-0002-5899-7523", "clpid": "Winfree-E" }, { "family_name": "Koch", "given_name": "Christof", "orcid": "0000-0001-6482-8067", "clpid": "Koch-C" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "Artificial neural networks are computationally powerful and exhibit brain-like dynamics. Unfortunately, the conventional gradient-dependent learning algorithms used to train them are biologically implausible. The calculation of the gradient in a traditional artificial neural network requires a complementary network of fast training signals that are dependent upon, but must not affect, the primary output-generating network activity. In contrast, the network of neurons in the cortex is highly recurrent; a network of gradient-calculating neurons in the brain would certainly project back to and influence the primary network. We address this biological implausibility by introducing a novel class of recurrent neural networks, intrinsic gradient networks, for which the gradient of an error function with respect to the parameters is a simple function of the network state. These networks can be trained using only their intrinsic signals, much like the network of neurons in the brain.
\r\n\r\nWe derive a simple equation that characterizes intrinsic gradient networks, and construct a broad set of networks that satisfy this characteristic equation. The resulting set of intrinsic gradient networks includes many highly recurrent instances for which the gradient can be calculated by a simple, local, pseudo-Hebbian function of the network state, thus resolving a long-standing contradiction between artificial and biological neural networks. We demonstrate that these networks can learn to perform nontrivial tasks like handwritten digit recognition using only their intrinsic signals. Finally, we show that a cortical implementation of an intrinsic gradient network would have a number of characteristic computational, anatomical, and electrophysiological properties, and review experimental evidence suggesting the manifestation of these properties in the cortex.
", "doi": "10.7907/YCB7-7X24", "publication_date": "2012", "thesis_type": "phd", "thesis_year": "2012" }, { "id": "thesis:6514", "collection": "thesis", "collection_id": "6514", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:06092011-140731576", "type": "thesis", "title": "Entropy Region and Network Information Theory", "author": [ { "family_name": "Shadbakht", "given_name": "Sormeh", "clpid": "Shadbakht-Sormeh" } ], "thesis_advisor": [ { "family_name": "Hassibi", "given_name": "Babak", "orcid": "0000-0002-1375-5838", "clpid": "Hassibi-B" } ], "thesis_committee": [ { "family_name": "Hassibi", "given_name": "Babak", "orcid": "0000-0002-1375-5838", "clpid": "Hassibi-B" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Ho", "given_name": "Tracey C.", "clpid": "Ho-Tracey" }, { "family_name": "Vaidyanathan", "given_name": "P. P.", "orcid": "0000-0003-3003-7042", "clpid": "Vaidyanathan-P-P" }, { "family_name": "Wierman", "given_name": "Adam C.", "orcid": "0000-0002-5923-0199", "clpid": "Wierman-A-C" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "This dissertation takes a step toward a general framework for solving network information theory problems by studying the capacity region of networks through the entropy region.
\r\n\r\nWe first show that the capacity of a large class of acyclic memoryless multiuser information theory problems can be formulated as convex optimization over the region of entropy vectors of the network random variables. This capacity characterization is universal, and is advantageous over previous formulations in that it is single letter. Besides, it is significant as it reveals the fundamental role of the entropy region in determining the capacity of network information theory problems.
\r\n\r\nWith this viewpoint, the rest of the thesis is dedicated to the study of the entropy region, and its consequences for networks. A full characterization of the entropy region has proven to be a very challenging problem, and thus, we mostly consider inner bound constructions. For discrete random variables, our approaches include characterization of entropy vectors with a lattice-derived probability distribution, the entropy region of binary random variables, and the linear representable region. Through these characterizations, and using matroid representability results, we study linear coding capacity of networks in different scenarios (e.g., binary operations in a network, or networks with two sources).
\r\n\r\nWe also consider continuous random variables by studying the entropy region of jointly Gaussian random variables. In particular, we determine the sufficiency of Gaussian random variables for characterizing the entropy region of 3 random variables in general. For more than 3 random variables, we point out the set of minimal necessary and sufficient conditions for a vector to be an entropy vector of jointly Gaussian random variables.
\r\n\r\nFinally, in the absence of a full analytical characterization of the entropy region, it is desirable to be able to perform numerical optimization over this space. In this regard, we propose a certain Monte Carlo method that enables one to numerically optimize entropy functions of discrete random variables, and also the achievable rates in wired networks. This method can be further adjusted for decentralized operation of networks. The promise of this technique is shown through various simulations of several interesting network problems.
\r\n", "doi": "10.7907/P8ZB-4D40", "publication_date": "2011", "thesis_type": "phd", "thesis_year": "2011" }, { "id": "thesis:6159", "collection": "thesis", "collection_id": "6159", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:10262010-082537998", "primary_object_url": { "basename": "submit_oct26.pdf", "content": "final", "filesize": 4906533, "license": "other", "mime_type": "application/pdf", "url": "/6159/1/submit_oct26.pdf", "version": "v5.0.0" }, "type": "thesis", "title": "SPICE\u00b2: A Spatial, Parallel Architecture for Accelerating the Spice Circuit Simulator\r ", "author": [ { "family_name": "Kapre", "given_name": "Nachiket Ganesh", "orcid": "0000-0002-2187-0406", "clpid": "Kapre-Nachiket-Ganesh" } ], "thesis_advisor": [ { "family_name": "DeHon", "given_name": "Andre", "clpid": "DeHon-A" } ], "thesis_committee": [ { "family_name": "Martin", "given_name": "Alain J.", "clpid": "Martin-A-J" }, { "family_name": "Meiron", "given_name": "Daniel I.", "clpid": "Meiron-D-I" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Trimberger", "given_name": "Steven", "clpid": "Trimberger-S" }, { "family_name": "DeHon", "given_name": "Andre", "clpid": "DeHon-A" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "Spatial processing of sparse, irregular floating-point computation using a single FPGA enables up to an order of magnitude speedup (mean 2.8X speedup) over a conventional microprocessor for the SPICE circuit simulator. We deliver this speedup using a hybrid parallel architecture that spatially implements the heterogeneous forms of parallelism available in SPICE. We decompose SPICE into its three constituent phases: Model-Evaluation, Sparse Matrix-Solve, and Iteration Control and parallelize each phase independently. We exploit data-parallel device evaluations in the Model-Evaluation phase, sparse dataflow parallelism in the Sparse Matrix-Solve phase and compose the complete design in streaming fashion. We name our parallel architecture SPICE\u00b2: Spatial Processors Interconnected for Concurrent Execution for accelerating the SPICE circuit simulator. We program the parallel architecture with a high-level, domain-specific framework that identifies, exposes and exploits parallelism available in the SPICE circuit simulator. This design is optimized with an auto-tuner that can scale the design to use larger FPGA capacities without expert intervention and can even target other parallel architectures with the assistance of automated code-generation. This FPGA architecture is able to outperform conventional processors due to a combination of factors including high utilization of statically-scheduled resources, low-overhead dataflow scheduling of fine-grained tasks, and overlapped processing of the control algorithms.
\r\n\r\nWe demonstrate that we can independently accelerate Model-Evaluation by a mean factor of 6.5X(1.4--23X) across a range of non-linear device models and Matrix-Solve by 2.4X(0.6--13X) across various benchmark matrices while delivering a mean combined speedup of 2.8X(0.2--11X) for the two together when comparing a Xilinx Virtex-6 LX760 (40nm) with an Intel Core i7 965 (45nm). With our high-level framework, we can also accelerate Single-Precision Model-Evaluation on NVIDIA GPUs, ATI GPUs, IBM Cell, and Sun Niagara 2 architectures.
\r\n\r\nWe expect approaches based on exploiting spatial parallelism to become important as frequency scaling slows down and modern processing architectures turn to parallelism (\\eg multi-core, GPUs) due to constraints of power consumption. This thesis shows how to express, exploit and optimize spatial parallelism for an important class of problems that are challenging to parallelize.
\r\n", "doi": "10.7907/QVZR-VB52", "publication_date": "2011", "thesis_type": "phd", "thesis_year": "2011" }, { "id": "thesis:5904", "collection": "thesis", "collection_id": "5904", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:06012010-161706139", "primary_object_url": { "basename": "Thesis_Sukwon_Kim_Caltech.pdf", "content": "final", "filesize": 1943455, "license": "other", "mime_type": "application/pdf", "url": "/5904/1/Thesis_Sukwon_Kim_Caltech.pdf", "version": "v6.0.0" }, "type": "thesis", "title": "Network Coding for Resource Optimization and Error Correction", "author": [ { "family_name": "Kim", "given_name": "Sukwon", "clpid": "Kim-Sukwon" } ], "thesis_advisor": [ { "family_name": "Effros", "given_name": "Michelle", "orcid": "0000-0003-3757-0675", "clpid": "Effros-M" }, { "family_name": "Ho", "given_name": "Tracey C.", "clpid": "Ho-Tracey" } ], "thesis_committee": [ { "family_name": "Effros", "given_name": "Michelle", "orcid": "0000-0003-3757-0675", "clpid": "Effros-M" }, { "family_name": "Ho", "given_name": "Tracey C.", "clpid": "Ho-Tracey" }, { "family_name": "Hassibi", "given_name": "Babak", "orcid": "0000-0002-1375-5838", "clpid": "Hassibi-B" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Low", "given_name": "Steven H.", "orcid": "0000-0001-6476-3048", "clpid": "Low-S-H" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "In the first part of this thesis, we demonstrate the benefits of network coding for optimizing the use of various network resources.
\r\n\r\nWe first study the problem of minimizing the power consumption for wireless multiple unicasts. A simple XOR-based coding strategy is considered for reducing the power consumption. We present a centralized polynomial time algorithm that approximately minimizes the number of transmissions for two unicast sessions. We extend it to a greedy algorithm for general problem of multiple unicasts.
\r\n\r\nPrevious results on network coding for low-power wireless transmissions of multiple unicasts rely on opportunistic coding or centralized optimization to reduce the power consumption. Thus we propose a distributed strategy for reducing the power consumption with wireless multiple unicasts. Our strategy attempts to increase network coding opportunities without the overhead required for centralized design or coordination. We present a polynomial time algorithm using our strategy that maximizes the expected power savings with respect to the random choice of sources and sinks on the wireless rectangular grid.
\r\n\r\nWe study the problem of minimum-energy multicast using network coding in mobile ad hoc networks (MANETs). The optimal subgraph can be obtained by solving a linear program every time slot, but it leads to high computational complexity. We present a low-complexity approach, network coding with periodic recomputation, which recomputes an approximate solution at fixed time intervals, and uses this solution during each time interval. We analyze the power consumption and the complexity of network with this approach.
\r\n\r\nRecently, several back-pressure type optimization algorithms with network coding are presented for multiple unicasts and multicast. Such algorithms are distributed since decisions are made locally at each node based on feedback about the size of queues at the destination node of each link. We develop a back-pressure based distributed optimization framework, which can be used for optimizing over any class of network codes. Our approach is to specify the class of coding operations by a set of generalized links, and to develop optimization tools that apply to any network composed of such generalized links.
\r\n\r\nIn the second part of this thesis, we study the capacity of single-source single-sink noiseless networks under adversarial attack on no more than z edges. Unlike prior papers, which assume equal capacities on all links, we allow arbitrary link capacities. Results include new upper bounds, general transmission strategies, and example networks where those bounds are tight. We introduce a new method for finding upper bounds on the linear coding capacities of arbitrary networks and show that there exists networks where the capacity is 50% greater than the best rate that can be achieved with linear coding. We also demonstrate examples where, unlike the equal link capacity case, it is necessary for intermediate nodes to do coding, nonlinear error detection or error correction in order to achieve the capacity. We introduce a new strategy called \"guess-and-forward\" and employ this strategy on a sequence of networks designed to provide increasingly complex generalizations of the cut-set bounds. The first is a two-node network with multiple feedback links. The second is a four-node acyclic network. The third is a family of 'zig-zag' networks. In the first two cases, the guess-and-forward strategy achieves the capacity. For zig-zag networks, we derive a achievable rate of guess-and-forward strategy.
", "doi": "10.7907/9E6W-SN15", "publication_date": "2010", "thesis_type": "phd", "thesis_year": "2010" }, { "id": "thesis:5904", "collection": "thesis", "collection_id": "5904", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:06012010-161706139", "primary_object_url": { "basename": "Thesis_Sukwon_Kim_Caltech.pdf", "content": "final", "filesize": 1943455, "license": "other", "mime_type": "application/pdf", "url": "/5904/1/Thesis_Sukwon_Kim_Caltech.pdf", "version": "v6.0.0" }, "type": "thesis", "title": "Network Coding for Resource Optimization and Error Correction", "author": [ { "family_name": "Kim", "given_name": "Sukwon", "clpid": "Kim-Sukwon" } ], "thesis_advisor": [ { "family_name": "Effros", "given_name": "Michelle", "orcid": "0000-0003-3757-0675", "clpid": "Effros-M" }, { "family_name": "Ho", "given_name": "Tracey C.", "clpid": "Ho-Tracey" } ], "thesis_committee": [ { "family_name": "Effros", "given_name": "Michelle", "orcid": "0000-0003-3757-0675", "clpid": "Effros-M" }, { "family_name": "Ho", "given_name": "Tracey C.", "clpid": "Ho-Tracey" }, { "family_name": "Hassibi", "given_name": "Babak", "orcid": "0000-0002-1375-5838", "clpid": "Hassibi-B" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Low", "given_name": "Steven H.", "orcid": "0000-0001-6476-3048", "clpid": "Low-S-H" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "In the first part of this thesis, we demonstrate the benefits of network coding for optimizing the use of various network resources.
\r\n\r\nWe first study the problem of minimizing the power consumption for wireless multiple unicasts. A simple XOR-based coding strategy is considered for reducing the power consumption. We present a centralized polynomial time algorithm that approximately minimizes the number of transmissions for two unicast sessions. We extend it to a greedy algorithm for general problem of multiple unicasts.
\r\n\r\nPrevious results on network coding for low-power wireless transmissions of multiple unicasts rely on opportunistic coding or centralized optimization to reduce the power consumption. Thus we propose a distributed strategy for reducing the power consumption with wireless multiple unicasts. Our strategy attempts to increase network coding opportunities without the overhead required for centralized design or coordination. We present a polynomial time algorithm using our strategy that maximizes the expected power savings with respect to the random choice of sources and sinks on the wireless rectangular grid.
\r\n\r\nWe study the problem of minimum-energy multicast using network coding in mobile ad hoc networks (MANETs). The optimal subgraph can be obtained by solving a linear program every time slot, but it leads to high computational complexity. We present a low-complexity approach, network coding with periodic recomputation, which recomputes an approximate solution at fixed time intervals, and uses this solution during each time interval. We analyze the power consumption and the complexity of network with this approach.
\r\n\r\nRecently, several back-pressure type optimization algorithms with network coding are presented for multiple unicasts and multicast. Such algorithms are distributed since decisions are made locally at each node based on feedback about the size of queues at the destination node of each link. We develop a back-pressure based distributed optimization framework, which can be used for optimizing over any class of network codes. Our approach is to specify the class of coding operations by a set of generalized links, and to develop optimization tools that apply to any network composed of such generalized links.
\r\n\r\nIn the second part of this thesis, we study the capacity of single-source single-sink noiseless networks under adversarial attack on no more than z edges. Unlike prior papers, which assume equal capacities on all links, we allow arbitrary link capacities. Results include new upper bounds, general transmission strategies, and example networks where those bounds are tight. We introduce a new method for finding upper bounds on the linear coding capacities of arbitrary networks and show that there exists networks where the capacity is 50% greater than the best rate that can be achieved with linear coding. We also demonstrate examples where, unlike the equal link capacity case, it is necessary for intermediate nodes to do coding, nonlinear error detection or error correction in order to achieve the capacity. We introduce a new strategy called \"guess-and-forward\" and employ this strategy on a sequence of networks designed to provide increasingly complex generalizations of the cut-set bounds. The first is a two-node network with multiple feedback links. The second is a four-node acyclic network. The third is a family of 'zig-zag' networks. In the first two cases, the guess-and-forward strategy achieves the capacity. For zig-zag networks, we derive a achievable rate of guess-and-forward strategy.
", "doi": "10.7907/9E6W-SN15", "publication_date": "2010", "thesis_type": "phd", "thesis_year": "2010" }, { "id": "thesis:5849", "collection": "thesis", "collection_id": "5849", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:05262010-173410602", "primary_object_url": { "basename": "DYZ_thesis.pdf", "content": "final", "filesize": 5915089, "license": "other", "mime_type": "application/pdf", "url": "/5849/1/DYZ_thesis.pdf", "version": "v4.0.0" }, "type": "thesis", "title": "Dynamic DNA Strand Displacement Circuits", "author": [ { "family_name": "Zhang", "given_name": "David Yu", "clpid": "Zhang-David-Yu" } ], "thesis_advisor": [ { "family_name": "Winfree", "given_name": "Erik", "clpid": "Winfree-E" } ], "thesis_committee": [ { "family_name": "Pierce", "given_name": "Niles A.", "clpid": "Pierce-N-A" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Fraser", "given_name": "Scott E.", "clpid": "Fraser-S-E" }, { "family_name": "Winfree", "given_name": "Erik", "clpid": "Winfree-E" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "Nucleic acids, the \"NA\" in DNA and RNA, have long been known to be vitally important molecules within biological cells and organisms. However, they are interesting for more than just their known roles in biology: their predictable Watson-Crick base pairing properties allow nucleic acids to be powerful nanoscale engineering tools. Additionally, nucleic acid-based devices are particularly attractive as biotechnological tools, because nucleic acids naturally exist within all life, and thus nucleic acid devices more easily function in cellular environments. It is for these reasons that nucleic acids have emerged as a frequent star in recent synthetic biology, biotechnology, and nanotechnology research papers.
\r\n\r\nThis thesis is a collection of 6 experimental papers, 3 theoretical papers, and 1 review paper that demonstrate and characterize novel nucleic acid-based devices such as catalysts, logic gates, and allosteric switches. Particular effort was placed in ensuring that all the designs are generalizable in sequence and that all the devices are modular in nature; this allows many different components to be integrated into higher-complexity devices.
\r\n\r\nThe works presented in this thesis were designed using only non-covalent changes to nucleic acid complexes and structures via Watson-Crick base pairing--i.e. hybridization, branch migration, and dissociation. These three primitives are sufficient to construct an endless variety of circuits and devices, much like how resistors, capacitors, and inductors allow complex electrical circuits. One advantage of devices, reactions, and circuits engineered using only Watson-Crick interactions is their robustness to their environmental conditions. While enzymatic reactions require specific temperatures, salt conditions, and co-factors, nucleic acid hybridization works reliably in a variety of different solutions.
\r\n\r\nThese works are not meant to be final, optimized designs for devices, but rather demonstrations of the wide range of possibilities afforded by nucleic acid engineering and of problems that can be practically solved with dynamic nucleic acid devices in the near future.
", "doi": "10.7907/Q4NR-CH87", "publication_date": "2010", "thesis_type": "phd", "thesis_year": "2010" }, { "id": "thesis:5848", "collection": "thesis", "collection_id": "5848", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:05262010-171937384", "primary_object_url": { "basename": "thesis.pdf", "content": "final", "filesize": 6183739, "license": "other", "mime_type": "application/pdf", "url": "/5848/1/thesis.pdf", "version": "v5.0.0" }, "type": "thesis", "title": "Transfer at C. elegans Synapses", "author": [ { "family_name": "Narayan", "given_name": "Anusha", "clpid": "Narayan-Anusha" } ], "thesis_advisor": [ { "family_name": "Sternberg", "given_name": "Paul W.", "orcid": "0000-0002-7699-0173", "clpid": "Sternberg-P-W" }, { "family_name": "Laurent", "given_name": "Gilles J.", "orcid": "0000-0002-2296-114X", "clpid": "Laurent-G-J" } ], "thesis_committee": [ { "family_name": "Siapas", "given_name": "Athanassios G.", "orcid": "0000-0001-8837-678X", "clpid": "Siapas-A-G" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Schuman", "given_name": "Erin Margaret", "orcid": "0000-0002-7053-1005", "clpid": "Schuman-E-M" }, { "family_name": "Sternberg", "given_name": "Paul W.", "orcid": "0000-0002-7699-0173", "clpid": "Sternberg-P-W" }, { "family_name": "Laurent", "given_name": "Gilles J.", "orcid": "0000-0002-2296-114X", "clpid": "Laurent-G-J" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "The nematode C.elegans, with its 302 neurons and abundance of genetic, laser ablation, electrophysiological and imaging tools, is a compact and attractive system for neural circuit analysis. An understanding of the functional dynamics of neural computation requires physiological analyses. We undertook the \ufb01rst characterization of transfer at central synapses in C.elegans. To achieve this we employed optical stimulation techniques using channelrhodopsin-2, and combined this with whole-cell patch clamp electrophysiological recording techniques. We show that the synapse between AFD and AIY, the \ufb01rst stage in the thermotactic circuit, exhibits excitatory, tonic and graded release. The gain at the synapse was low (<0.1), and release was frequency independent, showing no signs of facilitation or depression. The AFD-AIY synapse thus seems designed for robust and reliable transmission of a scaled-down temperature signal from AFD to AIY, enabling AIY to continuously monitor temperature information and integrate it with other incoming sensory information. We also investigated the synapse between ASER, a chemosensory neuron, and AIY, and found that the synaptic response was small and inconsistent. The combination of optical stimulation tools with neural recording techniques is a powerful way to analyze neural circuitry, and will be a signi\ufb01cant aid in achieving the goal of understanding how information is processed in the compact yet densely interconnected nervous system of the worm.", "doi": "10.7907/CB9Q-MC14", "publication_date": "2010", "thesis_type": "phd", "thesis_year": "2010" }, { "id": "thesis:5260", "collection": "thesis", "collection_id": "5260", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-11092007-180524", "type": "thesis", "title": "Soft-Error Tolerant Quasi Delay-insensitive Circuits", "author": [ { "family_name": "Jang", "given_name": "Wonjin", "clpid": "Jang-Wonjin" } ], "thesis_advisor": [ { "family_name": "Martin", "given_name": "Alain J.", "clpid": "Martin-A-J" } ], "thesis_committee": [ { "family_name": "Martin", "given_name": "Alain J.", "clpid": "Martin-A-J" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Chandy", "given_name": "K. Mani", "orcid": "0000-0001-9190-1290", "clpid": "Chandy-K-M" }, { "family_name": "Ho", "given_name": "Tracey C.", "clpid": "Ho-Tracey" }, { "family_name": "Abu-Mostafa", "given_name": "Yaser S.", "clpid": "Abu-Mostafa-Y-S" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "A hard error is an error that damages a circuit irrevocably; a soft error flips the logic states without causing any physical damage to the circuit, resulting in transient corruption of data. They result in transient, inconsistent corruption of data.
\r\n\r\nThe soft-error tolerance of logic circuits is recently getting more attention, since the soft- error rate of advanced CMOS devices is higher than before. As a response to the concern on soft errors, we propose a new method for making asynchronous circuits tolerant to soft errors. Since it relies on a property unique to asynchronous circuits, the method is different from what is done in synchronous circuits with triple modular redundancy. Asynchronous circuits have been attractive to the designers of reliable systems, because of their clock-less design, which makes them more robust to variations on computation time of modules. The quasi delay-insensitive (QDI) design style is one of the most robust asynchronous design styles for general computation; it makes one minimal assumption on delays in gates and wires. QDI circuits are easy to verify, simple, and modular, because the correct operation of a QDI circuit is independent of delays in gates and wires.
\r\n\r\nHere, we shall overview how to design a QDI circuit, and what will happen if a soft error occurs on a QDI circuit. Then the crucial components of the method are shown: (1) a special kind of duplication for random logic (when each bit has to be corrected individually), (2) special protection circuitry for arbiter and synchronizer (as needed for example for external interrupts), (3) reconfigurable circuits using a special configuration unit, and (4) error correcting for memory arrays and other structures in which the data bits can be self- corrected. The solution of protecting random logic is compared with alternatives, which use other types of error correcting codes (e.g., parity code) in a QDI circuit. It turns out that the duplication generates efficient circuits more commonly than other possible constructions. Finally, the design of a soft-error tolerant asynchronous microprocessor is detailed and testing results of the soft-error tolerance of the microprocessor are shown.
\r\n\r\n", "doi": "10.7907/ZVFF-WE07", "publication_date": "2008", "thesis_type": "phd", "thesis_year": "2008" }, { "id": "thesis:2261", "collection": "thesis", "collection_id": "2261", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-05292008-142339", "primary_object_url": { "basename": "thesis.pdf", "content": "final", "filesize": 2528018, "license": "other", "mime_type": "application/pdf", "url": "/2261/1/thesis.pdf", "version": "v3.0.0" }, "type": "thesis", "title": "Molecules Computing: Self-Assembled Nanostructures, Molecular Automata,and Chemical Reaction Networks", "author": [ { "family_name": "Soloveichik", "given_name": "David", "orcid": "0000-0002-2585-4120", "clpid": "Soloveichik-David" } ], "thesis_advisor": [ { "family_name": "Winfree", "given_name": "Erik", "clpid": "Winfree-E" } ], "thesis_committee": [ { "family_name": "Abu-Mostafa", "given_name": "Yaser S.", "clpid": "Abu-Mostafa-Y-S" }, { "family_name": "Umans", "given_name": "Christopher M.", "clpid": "Umans-C-M" }, { "family_name": "Winfree", "given_name": "Erik", "clpid": "Winfree-E" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Doyle", "given_name": "John Comstock", "clpid": "Doyle-J-C" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "Many endeavors of molecular-level engineering either rely on biological material such as nucleic acids and restriction enzymes, or are inspired by biological processes such as self-assembly or cellular regulatory networks. This thesis develops theories on three such topics: self-assembled nanostructures, molecular automata, and chemical reaction networks. The abstractions and underlying methods of the theories presented herein are based on computer science and include Turing machines and circuits. Toward engineering self-assembled nanostructures, we create a theory of scale-free shapes in which the complexity of their self-assembly is connected to the shapes' descriptional complexity. Further, we study patterns in terms of whether they can be self-assembled robustly without an increase in scale to accommodate redundancy. We also describe a new method of ensuring resilience to more types of error simultaneously. Toward creating molecular automata we study the computational power of a restriction enzyme-based automaton. Toward designing chemical reaction networks, we develop a technique of storing and processing information in molecular counts, which is capable of achieving Turing universal computation. We also study the computational complexity of simulating stochastic chemical reaction networks and formally connect robustness and simulation efficiency. Lastly, we describe nucleic acid implementations of Boolean logic circuits and arbitrary mass-action kinetics. The three areas of this thesis are promising realizations of molecular-level engineering, and the theories presented here inform the range of possibility or delineate inherent difficulties in these areas. ", "doi": "10.7907/ZGE0-AF38", "publication_date": "2008", "thesis_type": "phd", "thesis_year": "2008" }, { "id": "thesis:2267", "collection": "thesis", "collection_id": "2267", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-05292008-231048", "primary_object_url": { "basename": "thesis.pdf", "content": "final", "filesize": 1612108, "license": "other", "mime_type": "application/pdf", "url": "/2267/1/thesis.pdf", "version": "v4.0.0" }, "type": "thesis", "title": "Adaptive Learning Algorithms and Data Cloning", "author": [ { "family_name": "Pratap", "given_name": "Amrit", "clpid": "Pratap-Amrit" } ], "thesis_advisor": [ { "family_name": "Abu-Mostafa", "given_name": "Yaser S.", "clpid": "Abu-Mostafa-Y-S" } ], "thesis_committee": [ { "family_name": "Abu-Mostafa", "given_name": "Yaser S.", "clpid": "Abu-Mostafa-Y-S" }, { "family_name": "Martin", "given_name": "Alain J.", "clpid": "Martin-A-J" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Perona", "given_name": "Pietro", "clpid": "Perona-P" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "This thesis is in the field of machine learning: the use of data to automatically learn a hypothesis to predict the future behavior of a system. It summarizes three of my research projects.
\r\n\r\nWe first investigate the role of margins in the phenomenal success of the Boosting Algorithms. AdaBoost (Adaptive Boosting) is an algorithm for generating an ensemble of hypotheses for classification. The superior out-of-sample performance of AdaBoost has been attributed to the fact that it can generate a classifier which classifies the points with a large margin of confidence. This led to the development of many new algorithms focusing on optimizing the margin of confidence. It was observed that directly optimizing the margins leads to a poor performance. This apparent contradiction has been the topic of a long unresolved debate in the machine-learning community. We introduce new algorithms which are expressly designed to test the margin hypothesis and provide concrete evidence which refutes the margin argument.
\r\n\r\nWe then propose a novel algorithm for Adaptive sampling under Monotonicity constraint. The typical learning problem takes examples of the target function as input information and produces a hypothesis that approximates the target as an output. We consider a generalization of this paradigm by taking different types of information as input, and producing only specific properties of the target as output. This is a very common setup which occurs in many different real-life settings where the samples are expensive to obtain. We show experimentally that our algorithm achieves better performance than the existing methods, such as Staircase procedure and PEST.
\r\n\r\nOne of the major pitfalls in machine learning research is that of selection bias. This is mostly introduced unconsciously due to the choices made during the learning process, which often lead to over-optimistic estimates of the performance. In the third project, we introduce a new methodology for systematically reducing selection bias. Experiments show that using cloned datasets for model selection can lead to better performance and reduce the selection bias.
", "doi": "10.7907/GV3D-AB69", "publication_date": "2008", "thesis_type": "phd", "thesis_year": "2008" }, { "id": "thesis:2313", "collection": "thesis", "collection_id": "2313", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-05302008-002429", "primary_object_url": { "basename": "thesis_double.pdf", "content": "final", "filesize": 2740336, "license": "other", "mime_type": "application/pdf", "url": "/2313/1/thesis_double.pdf", "version": "v4.0.0" }, "type": "thesis", "title": "Generalized Network Routing Metrics and Algorithms", "author": [ { "family_name": "Soedarmadji", "given_name": "Edwin", "clpid": "Soedarmadji-Edwin" } ], "thesis_advisor": [ { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" } ], "thesis_committee": [ { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" }, { "family_name": "Hassibi", "given_name": "Babak", "orcid": "0000-0002-1375-5838", "clpid": "Hassibi-B" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Low", "given_name": "Steven H.", "orcid": "0000-0001-6476-3048", "clpid": "Low-S-H" }, { "family_name": "Ho", "given_name": "Tracey C.", "clpid": "Ho-Tracey" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "In this thesis, we introduce generalized network routing metrics that represent probability density parameters of the most popular communication channel models such as (a) the q-ary Symmetric Channel (q-SC) (b) the q-ary Erasure Channel (q-EC); (c) the Gilbert-Elliot Channel (GEC); and (d) the constrained Additive White Gaussian Noise (AWGN). The GEC is a very important for modelling correlated errors in channels such as the ubiquitous TCP/IP links and the wireless fading channels. In this thesis, we prove that channel models (a)--(d) can be used as inputs to the Generalized Dijkstra's Algorithm without resulting in any routing loop.
\r\n\r\nWe also define our own generalized Dijkstra's algorithm that can solve a modified standard shortest path problem that features: (1) a subset of network nodes that are capable of reducing the accumulated path cost down to zero, and (2) a constraint that the cumulative cost of any feasible path must never exceed a prespecified maximum value. We call this modified problem the Gas Station Problem, and its solution the Gas Station Algorithm. The algorithm can be applied in many different areas such as: vehicle routing, project management, and most importantly, network communication.
\r\n\r\nWe investigate various auxilliary synchronization algorithms used in popular routing protocols. Synchronization is used by routers to ensure that all routers operate on an identical routing table --- not a trivial task, considering network unreliabilities and possible malicious attacks. Our analysis produces a list of assumptions that guarantees synchronization. We also obtain the upper bounds to quantities such as transmission period, memory requirement, etc. In turn, these bounds can be used to rate network performance.
\r\n\r\nFinally, in a related contribution, we analyze message synchronization where a message is retransmitted only if the number of identical messages received exceeds a certain threshold. We define the Chinese Generals Problem as the problem of identifying the set of assumptions under which synchronization is guaranteed. This threshold-base message passing algorithm has the benefits of a tunable gain and a higher noise resistance.
\r\n", "doi": "10.7907/Z94B2ZJD", "publication_date": "2008", "thesis_type": "phd", "thesis_year": "2008" }, { "id": "thesis:2313", "collection": "thesis", "collection_id": "2313", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-05302008-002429", "primary_object_url": { "basename": "thesis_double.pdf", "content": "final", "filesize": 2740336, "license": "other", "mime_type": "application/pdf", "url": "/2313/1/thesis_double.pdf", "version": "v4.0.0" }, "type": "thesis", "title": "Generalized Network Routing Metrics and Algorithms", "author": [ { "family_name": "Soedarmadji", "given_name": "Edwin", "clpid": "Soedarmadji-Edwin" } ], "thesis_advisor": [ { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" } ], "thesis_committee": [ { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" }, { "family_name": "Hassibi", "given_name": "Babak", "orcid": "0000-0002-1375-5838", "clpid": "Hassibi-B" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Low", "given_name": "Steven H.", "orcid": "0000-0001-6476-3048", "clpid": "Low-S-H" }, { "family_name": "Ho", "given_name": "Tracey C.", "clpid": "Ho-Tracey" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "In this thesis, we introduce generalized network routing metrics that represent probability density parameters of the most popular communication channel models such as (a) the q-ary Symmetric Channel (q-SC) (b) the q-ary Erasure Channel (q-EC); (c) the Gilbert-Elliot Channel (GEC); and (d) the constrained Additive White Gaussian Noise (AWGN). The GEC is a very important for modelling correlated errors in channels such as the ubiquitous TCP/IP links and the wireless fading channels. In this thesis, we prove that channel models (a)--(d) can be used as inputs to the Generalized Dijkstra's Algorithm without resulting in any routing loop.
\r\n\r\nWe also define our own generalized Dijkstra's algorithm that can solve a modified standard shortest path problem that features: (1) a subset of network nodes that are capable of reducing the accumulated path cost down to zero, and (2) a constraint that the cumulative cost of any feasible path must never exceed a prespecified maximum value. We call this modified problem the Gas Station Problem, and its solution the Gas Station Algorithm. The algorithm can be applied in many different areas such as: vehicle routing, project management, and most importantly, network communication.
\r\n\r\nWe investigate various auxilliary synchronization algorithms used in popular routing protocols. Synchronization is used by routers to ensure that all routers operate on an identical routing table --- not a trivial task, considering network unreliabilities and possible malicious attacks. Our analysis produces a list of assumptions that guarantees synchronization. We also obtain the upper bounds to quantities such as transmission period, memory requirement, etc. In turn, these bounds can be used to rate network performance.
\r\n\r\nFinally, in a related contribution, we analyze message synchronization where a message is retransmitted only if the number of identical messages received exceeds a certain threshold. We define the Chinese Generals Problem as the problem of identifying the set of assumptions under which synchronization is guaranteed. This threshold-base message passing algorithm has the benefits of a tunable gain and a higher noise resistance.
\r\n", "doi": "10.7907/Z94B2ZJD", "publication_date": "2008", "thesis_type": "phd", "thesis_year": "2008" }, { "id": "thesis:2321", "collection": "thesis", "collection_id": "2321", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-05302008-143505", "primary_object_url": { "basename": "phdthesis.pdf", "content": "final", "filesize": 847038, "license": "other", "mime_type": "application/pdf", "url": "/2321/1/phdthesis.pdf", "version": "v2.0.0" }, "type": "thesis", "title": "From Ordinal Ranking to Binary Classification", "author": [ { "family_name": "Lin", "given_name": "Hsuan-Tien", "orcid": "0000-0003-2968-0671", "clpid": "Lin-Hsuan-Tien" } ], "thesis_advisor": [ { "family_name": "Abu-Mostafa", "given_name": "Yaser S.", "clpid": "Abu-Mostafa-Y-S" } ], "thesis_committee": [ { "family_name": "Abu-Mostafa", "given_name": "Yaser S.", "clpid": "Abu-Mostafa-Y-S" }, { "family_name": "Umans", "given_name": "Christopher M.", "clpid": "Umans-C-M" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Perona", "given_name": "Pietro", "clpid": "Perona-P" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "We study the ordinal ranking problem in machine learning. The problem can be viewed as a classification problem with additional ordinal information or as a regression problem without actual numerical information. From the classification perspective, we formalize the concept of ordinal information by a cost-sensitive setup, and propose some novel cost-sensitive classification algorithms. The algorithms are derived from a systematic cost-transformation technique, which carries a strong theoretical guarantee. Experimental results show that the novel algorithms perform well both in a general cost-sensitive setup and in the specific ordinal ranking setup.
\r\n\r\nFrom the regression perspective, we propose the threshold ensemble model for ordinal ranking, which allows the machines to estimate a real-valued score (like regression) before quantizing it to an ordinal rank. We study the generalization ability of threshold ensembles and derive novel large-margin bounds on its expected test performance. In addition, we improve an existing algorithm and propose a novel algorithm for constructing large-margin threshold ensembles. Our proposed algorithms are efficient in training and achieve decent out-of-sample performance when compared with the state-of-the-art algorithm on benchmark data sets.
\r\n\r\nWe then study how ordinal ranking can be reduced to weighted binary classification. The reduction framework is simpler than the cost-sensitive classification approach and includes the threshold ensemble model as a special case. The framework allows us to derive strong theoretical results that tightly connect ordinal ranking with binary classification. We demonstrate the algorithmic and theoretical use of the reduction framework by extending SVM and AdaBoost, two of the most popular binary classification algorithms, to the area of ordinal ranking. Coupling SVM with the reduction framework results in a novel and faster algorithm for ordinal ranking with superior performance on real-world data sets, as well as a new bound on the expected test performance for generalized linear ordinal rankers. Coupling AdaBoost with the reduction framework leads to a novel algorithm that boosts the training accuracy of any cost-sensitive ordinal ranking algorithms theoretically, and in turn improves their test performance empirically.
\r\n\r\nFrom the studies above, the key to improve ordinal ranking is to improve binary classification. In the final part of the thesis, we include two projects that aim at understanding binary classification better in the context of ensemble learning. First, we discuss how AdaBoost is restricted to combining only a finite number of hypotheses and remove the restriction by formulating a framework of infinite ensemble learning based on SVM. The framework can output an infinite ensemble through embedding infinitely many hypotheses into an~SVM kernel. Using the framework, we show that binary classification (and hence ordinal ranking) can be improved by going from a finite ensemble to an infinite one. Second, we discuss how AdaBoost carries the property of being resistant to overfitting. Then, we propose the SeedBoost algorithm, which uses the property as a machinery to prevent other learning algorithms from overfitting. Empirical results demonstrate that SeedBoost can indeed improve an overfitting algorithm on some data sets.
", "doi": "10.7907/7B0F-E145", "publication_date": "2008", "thesis_type": "phd", "thesis_year": "2008" }, { "id": "thesis:328", "collection": "thesis", "collection_id": "328", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-01242008-134459", "primary_object_url": { "basename": "thesis_ycassuto.pdf", "content": "final", "filesize": 1124634, "license": "other", "mime_type": "application/pdf", "url": "/328/1/thesis_ycassuto.pdf", "version": "v3.0.0" }, "type": "thesis", "title": "Coding Techniques for Data-Storage Systems", "author": [ { "family_name": "Cassuto", "given_name": "Yuval", "orcid": "0000-0001-6369-6699", "clpid": "Cassuto-Yuval" } ], "thesis_advisor": [ { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" } ], "thesis_committee": [ { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Xu", "given_name": "Lihao", "clpid": "Xu-Lihao" }, { "family_name": "Blaum", "given_name": "Mario", "clpid": "Blaum-M" }, { "family_name": "Farrell", "given_name": "Paddy", "clpid": "Farrell-P" }, { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" }, { "family_name": "Ho", "given_name": "Tracey C.", "clpid": "Ho-Tracey" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "As information-bearing objects, data-storage systems are natural consumers of information-theoretic ideas. For many issues in data-storage systems, the best trade-off between cost, performance and reliability, passes through the application of error-correcting codes. Error-correcting codes that are specialized for data-storage systems is the subject studied by this thesis. On the practical side, central challenges of storage systems are addressed, both at the individual-device level and higher at the enterprise level for disk arrays. The results for individual devices include a new coding paradigm for Multi-Level Flash storage that benefits storage density and access speed, and also a higher-throughput algorithm for decoding Reed-Solomon codes with large decoding radii. The results for storage arrays address models and constructions to combat correlated device failures, and also introduce new highly-regular array-code constructions with optimal redundancy and updates. On the theoretical side, the research stretches across multiple layers of coding theory innovation: new codes for new error models, new codes for existing error models, and new decoding techniques for known codes. To bridge the properties and constraints of practical systems with the mathematical language of coding theory, new well-motivated models and abstractions are proposed. Among them are the models of t asymmetric limited-magnitude errors and clustered erasures. Later, after maximizing the theory's power in addressing the abstractions, the performance of storage systems that employ the new schemes is analytically validated.", "doi": "10.7907/38ZT-QT95", "publication_date": "2008", "thesis_type": "phd", "thesis_year": "2008" }, { "id": "thesis:4747", "collection": "thesis", "collection_id": "4747", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-12032007-113628", "primary_object_url": { "basename": "thesisETD.pdf", "content": "final", "filesize": 1011757, "license": "other", "mime_type": "application/pdf", "url": "/4747/1/thesisETD.pdf", "version": "v3.0.0" }, "type": "thesis", "title": "Optimization Algorithms in Wireless and Quantum Communications", "author": [ { "family_name": "Stojnic", "given_name": "Mihailo", "clpid": "Stojnic-Mihailo" } ], "thesis_advisor": [ { "family_name": "Hassibi", "given_name": "Babak", "orcid": "0000-0002-1375-5838", "clpid": "Hassibi-B" } ], "thesis_committee": [ { "family_name": "Hassibi", "given_name": "Babak", "orcid": "0000-0002-1375-5838", "clpid": "Hassibi-B" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Vaidyanathan", "given_name": "P. P.", "orcid": "0000-0003-3003-7042", "clpid": "Vaidyanathan-P-P" }, { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" }, { "family_name": "Ho", "given_name": "Tracey C.", "clpid": "Ho-Tracey" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "Since the first communication systems were developed, the scientific community has been witnessing attempts to increase the amount of information that can be transmitted. In the last 10--15 years there has been a tremendous amount of research towards developing multi-antenna systems which would hopefully provide high-data-rate transmissions. However, increasing the overall amount of transmitted information increases the complexity of the necessary signal processing. A large portion of this thesis deals with several important issues in signal processing of multi-antenna systems. In almost every particular case the goal is to develop a technique/algorithm so that the overall complexity of the signal processing is significantly decreased.
\r\n\r\nIn the first part of the thesis a very important problem of signal detection in MIMO (multiple-input multiple-output) systems is considered. The problem is analyzed in two different scenarios: when the transmission medium (channel) 1) is known and 2) is unknown at the receiver. The former case is often called coherent and the later non-coherent MIMO detection. Both cases usually amount to solving highly complex NP-hard combinatorial optimization problems. For the coherent case we develop a significant improvement of the traditional sphere decoder algorithm commonly used for this type of detection. An interesting connection between the new improved algorithm and the H-infinity estimation theory is established, and the performance improvement over the standard sphere decoder is demonstrated. For the non-coherent case we develop a counterpart to the standard sphere decoder, the so-called out-sphere decoder. The complexity of the algorithm is viewed as a random variable; its expected value is analyzed and shown to be significantly smaller than the one of the overall exhaustive search. In the non-coherent case, in addition to the complexity analysis of the exact out-sphere decoder, we analyze the performance loss of a suboptimal technique. We show that only a moderate loss of a few dbs in power required at the transmitter will occur if a polynomial algorithm based on the semi-definite relaxation is used in place of any exact technique (which of course is not known to be polynomial).
\r\n\r\nIn the second part of the thesis we consider a few problems that arise in wireless broadcast channels. Namely, we consider the problem of the information symbol vector design at the transmitter. A polynomial linear precoding technique is constructed. It enables achieving data rates very close to the ones achieved with DPC (dirty paper coding) technique. Additionally, for another suboptimal polynomial scheme (the so-called nulling and cancelling), we show that it asymptotically achieves the same data rate as the optimal, exponentially complex, DPC.
\r\n\r\nIn the last part of the thesis we consider a quantum counterpart of the signal detection from classical communication. In quantum systems the signals are quantum states and the quantum detection problem amounts to designing measurement operators which have to satisfy certain quantum mechanics laws. A specific type of quantum detection called unambiguous detection, which has numerous applications including quantum filtering, has recently attracted a lot of attention in the research community. We develop a general framework for numerically solving this problem using the tools from the convex optimization theory. Furthermore, in the special case where the two quantum states are of rank 2, we construct an explicit analytical solution for the measurement operators.
\r\n\r\nAt the end we would like to emphasize that the contribution of this thesis goes beyond the specific problems mentioned here. Most algorithmic optimization techniques developed in this paper are generally applicable. While it is a fact that our results were originally motivated by wireless and quantum communications applications, we believe that the developed techniques will find applications in many different areas where similar optimization problems appear.
\r\n", "doi": "10.7907/D6RN-ZD88", "publication_date": "2008", "thesis_type": "phd", "thesis_year": "2008" }, { "id": "thesis:5260", "collection": "thesis", "collection_id": "5260", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-11092007-180524", "type": "thesis", "title": "Soft-Error Tolerant Quasi Delay-insensitive Circuits", "author": [ { "family_name": "Jang", "given_name": "Wonjin", "clpid": "Jang-Wonjin" } ], "thesis_advisor": [ { "family_name": "Martin", "given_name": "Alain J.", "clpid": "Martin-A-J" } ], "thesis_committee": [ { "family_name": "Martin", "given_name": "Alain J.", "clpid": "Martin-A-J" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Chandy", "given_name": "K. Mani", "orcid": "0000-0001-9190-1290", "clpid": "Chandy-K-M" }, { "family_name": "Ho", "given_name": "Tracey C.", "clpid": "Ho-Tracey" }, { "family_name": "Abu-Mostafa", "given_name": "Yaser S.", "clpid": "Abu-Mostafa-Y-S" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "A hard error is an error that damages a circuit irrevocably; a soft error flips the logic states without causing any physical damage to the circuit, resulting in transient corruption of data. They result in transient, inconsistent corruption of data.
\r\n\r\nThe soft-error tolerance of logic circuits is recently getting more attention, since the soft- error rate of advanced CMOS devices is higher than before. As a response to the concern on soft errors, we propose a new method for making asynchronous circuits tolerant to soft errors. Since it relies on a property unique to asynchronous circuits, the method is different from what is done in synchronous circuits with triple modular redundancy. Asynchronous circuits have been attractive to the designers of reliable systems, because of their clock-less design, which makes them more robust to variations on computation time of modules. The quasi delay-insensitive (QDI) design style is one of the most robust asynchronous design styles for general computation; it makes one minimal assumption on delays in gates and wires. QDI circuits are easy to verify, simple, and modular, because the correct operation of a QDI circuit is independent of delays in gates and wires.
\r\n\r\nHere, we shall overview how to design a QDI circuit, and what will happen if a soft error occurs on a QDI circuit. Then the crucial components of the method are shown: (1) a special kind of duplication for random logic (when each bit has to be corrected individually), (2) special protection circuitry for arbiter and synchronizer (as needed for example for external interrupts), (3) reconfigurable circuits using a special configuration unit, and (4) error correcting for memory arrays and other structures in which the data bits can be self- corrected. The solution of protecting random logic is compared with alternatives, which use other types of error correcting codes (e.g., parity code) in a QDI circuit. It turns out that the duplication generates efficient circuits more commonly than other possible constructions. Finally, the design of a soft-error tolerant asynchronous microprocessor is detailed and testing results of the soft-error tolerance of the microprocessor are shown.
\r\n\r\n", "doi": "10.7907/ZVFF-WE07", "publication_date": "2008", "thesis_type": "phd", "thesis_year": "2008" }, { "id": "thesis:1561", "collection": "thesis", "collection_id": "1561", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-05012007-133654", "primary_object_url": { "basename": "Cedric_Anen_PhD_Thesis.pdf", "content": "final", "filesize": 5260880, "license": "other", "mime_type": "application/pdf", "url": "/1561/1/Cedric_Anen_PhD_Thesis.pdf", "version": "v2.0.0" }, "type": "thesis", "title": "Neural Correlates of Economic and Moral Decision-Making", "author": [ { "family_name": "Anen", "given_name": "C\u00e9dric Robert", "clpid": "Anen-C\u00e9dric-Robert" } ], "thesis_advisor": [ { "family_name": "Quartz", "given_name": "Steven R.", "clpid": "Quartz-S-R" } ], "thesis_committee": [ { "family_name": "Quartz", "given_name": "Steven R.", "clpid": "Quartz-S-R" }, { "family_name": "Camerer", "given_name": "Colin F.", "clpid": "Camerer-C-F" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Vaidyanathan", "given_name": "P. P.", "clpid": "Vaidyanathan-P-P" }, { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "Our daily lives are shaped by a series of decision processes, ranging from very unimportant choices to life-changing judgments. The complexity of the decision processes increases tremendously when the decision-making takes place in a social context, i.e., when other human beings are directly involved in the decision. In such conditions the decision-maker not only tries to maximize his own utility, but also needs to take into account the interdependent nature of the situation. Information about others' preferences, characteristics, and actions play an important role, and need to be thoroughly evaluated and predicted before making a decision. In this thesis we explore the neural correlates of two different types of social decision-making.
\r\n\r\nIn the first experiment we investigate economic decision-making in the context of a two-player social exchange game. In order to maximize their overall and personal earnings, players need to cooperate and build up a trust relationship with their partner. Synchronized neural data is recorded from the two interacting brains using functional magnetic resonance imaging. In this thesis we present four main findings: (i) the neural correlates of strategic uncertainty and how it can be used to predict a player's future strategic choice; (ii) the dynamic interaction of the brains of two interacting players; (iii) the neural correlates of trust and its development over the course of the game; and (iv) how the brain distinguishes between one's own actions and those of another person.
\r\n\r\nThe second experiment investigates the neural basis of moral decision-making and other- regarding preferences. Subjects have to make a morally difficult decision between helping two groups of children while trading off between efficiency and equity. By parametrically varying these variables, we show how two brain structures, the insula and the caudate, are actively involved in the decision-making process.
\r\n\r\nTaken together the results presented in this thesis shed some light on how our brain evaluates social situations, and how it uses social measures such as trust, agency, strategic interaction, and fairness to make decisions.
\r\n", "doi": "10.7907/W3VS-MW78", "publication_date": "2007", "thesis_type": "phd", "thesis_year": "2007" }, { "id": "thesis:4413", "collection": "thesis", "collection_id": "4413", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-11052006-173021", "primary_object_url": { "basename": "thesis_final.pdf", "content": "final", "filesize": 1820220, "license": "other", "mime_type": "application/pdf", "url": "/4413/1/thesis_final.pdf", "version": "v2.0.0" }, "type": "thesis", "title": "Performance Limits and Design Issues in Wireless Networks", "author": [ { "family_name": "Farajidana", "given_name": "Amir", "clpid": "Farajidana-Amir" } ], "thesis_advisor": [ { "family_name": "Hassibi", "given_name": "Babak", "orcid": "0000-0002-1375-5838", "clpid": "Hassibi-B" } ], "thesis_committee": [ { "family_name": "Hassibi", "given_name": "Babak", "orcid": "0000-0002-1375-5838", "clpid": "Hassibi-B" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Vaidyanathan", "given_name": "P. P.", "orcid": "0000-0003-3003-7042", "clpid": "Vaidyanathan-P-P" }, { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" }, { "family_name": "Ho", "given_name": "Tracey C.", "clpid": "Ho-Tracey" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "The increasing utilization of networks, especially wireless networks, for different applications and in different aspects of modern life, has directed a great deal of attention towards the analysis and optimal design of networks. Distinguishing features of the wireless environment and the distributed nature of the network setup have raised many important challenges in finding the performance limits of different tasks such as communication, control, and computation over networks. There are also many design issues concerning the complexity and the robustness of wireless systems that should be addressed for a thorough understanding and an efficient operation of wireless networked systems. This thesis deals with a few of the challenges associated with the fundamental performance limits and optimal design of wireless networks.
\r\n\r\nIn the first part, we analyze performance limits of two applications for a special class of wireless networks called wireless erasure networks. These networks incorporate some of the essential features of the wireless environment. We look at the performance limits of two applications over these networks. The first application is data transmission with two different traffic patterns, namely multicast and broadcast. The capacity region and the optimal coding scheme for the multicast scenario are found, and outer and inner bounds on the capacity region for the broadcast scenario are provided. The second application considered in this thesis is estimation and control of a dynamical process at a remote location connected through a wireless erasure network to a sensor observing the process. In this case, we characterize the minimum steady-state error and its dependency on the parameters of the network. The final problem considered in the first part of the thesis concerns power consumption (as a performance measure) in wireless networks. We propose and analyze a simple scheme based on the idea of distributed beamforming that saves us in terms of power consumption for dense sensor and ad-hoc networks. We quantify this gain compared to the case when nodes have isolated communications without participating in the network.
\r\n\r\nThe second part of the thesis deals with two design issues in the downlink of cellular wireless networks. The first issue is related to quality of service provisioning in the downlink scenario. We investigate the problem of differentiated rate scheduling in which different users demand different sets of rates. We obtain explicit and practical scheduling schemes to achieve the rate constraints and at the same time maximize the throughput. These schemes are based on the idea of opportunistic beamforming, are simple, and require little amount of feedback to the transmitter. We further show that the throughput loss due to imposing the rate constraints is negligible for large systems.
\r\n\r\nThe next issue considered in this thesis is the robustness of the capacity region of multiple antenna Gaussian broadcast channels to the channel estimation error at the transmitter and the users. These channels are mathematical models for the downlink of cellular systems. We provide an inner bound on the capacity region of these channels and show that this inner bound is equivalent to the capacity region of a dual multiple access channel with a noise covariance that depends on the transmit powers. This duality is explored to show the effect of the estimation error on the sum-rate for a large number of users and in the large power regime. Finally, a training-based scheme for the block fading multiple antenna broadcast channels is proposed.
", "doi": "10.7907/PRMQ-0644", "publication_date": "2007", "thesis_type": "phd", "thesis_year": "2007" }, { "id": "thesis:2203", "collection": "thesis", "collection_id": "2203", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-05282007-203227", "primary_object_url": { "basename": "Thesis_Komijani.pdf", "content": "final", "filesize": 9579874, "license": "other", "mime_type": "application/pdf", "url": "/2203/1/Thesis_Komijani.pdf", "version": "v2.0.0" }, "type": "thesis", "title": "Microwave Integrated Phased-Array Transmitters in Silicon", "author": [ { "family_name": "Komijani", "given_name": "Abbas", "clpid": "Komijani-Abbas" } ], "thesis_advisor": [ { "family_name": "Hajimiri", "given_name": "Ali", "clpid": "Hajimiri-A" } ], "thesis_committee": [ { "family_name": "Hajimiri", "given_name": "Ali", "clpid": "Hajimiri-A" }, { "family_name": "Yang", "given_name": "Changhuei", "clpid": "Yang-Changhuei" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "D'Addario", "given_name": "Larry R.", "clpid": "D'Addario-L-R" }, { "family_name": "Weinreb", "given_name": "Sander", "clpid": "Weinreb-S" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "Phased-array systems, a special case of multiple-input-multiple-output (MIMO) systems, take advantage of spatial directivity and array gain to increase spectral efficiency. Implementing a phased-array system at high frequency in a commercial silicon process technology presents several challenges. This thesis focuses on the architectural and circuit-level trade-offs involved in the design of the silicon-based fully integrated phased-array transmitters.
\r\n \r\nAs the first implementation, a four-element 24GHz 0.18\u00b5m CMOS phased-array transmitter with integrated power amplifiers is presented. On-chip power amplifiers use substrate-shielded slow-wave transmission lines for impedance matching and can generate up to 14dBm of output power. The transmitter employs a two-step upconversion architecture with 4.8GHz as the intermediate frequency (IF) and uses a single 19.2GHz synthesizer serving as the local oscillator (LO) generator. The phased-array, employing the LO phase shifting architecture, achieves 23dB of peak to null-ratio when all four elements are used, demonstrates a beam steering range covering all signal incident angles, and can support a data rate of 500Mbps with a quadrature phase-shift keying (QPSK) baseband signal.
\r\n\r\nAs the second implementation with a modified phase shifting architecture, an integrated 4-element 77GHz Silicon-Germanium (SiGe) phased-array transceiver is presented. Two-step conversion, envisioning a dual-mode 77GHz/24GHz operation, is used at both the receiver and the transmitter paths. A differential phase of 52GHz is generated by the on-chip voltage-controlled oscillator (VCO) and is distributed to all radio frequency (RF) paths. The phase shifting is performed at the LO ports of the RF mixers with continuous analog phase shifters. The quadrature signal of the second LO, at the IF frequency of 26GHz, is generated by dividing the VCO frequency by a factor of 2 using a cross-coupled injection-locked frequency divider. The on-chip 77GHz power amplifier with an output power of 17.5dBm and peak power added efficiency (PAE) of 14% achieves the best performance demonstrated in silicon. A single transmitter path achieves a 40dB conversion gain at 77GHz with 2.5GHz of bandwidth and a maximum output power of 12.5 dBm.
\r\n\r\nThe measured results demonstrate the feasibility of using silicon-based integrated phased-arrays for wireless communication and vehicular radar applications.
", "doi": "10.7907/EW67-RX66", "publication_date": "2007", "thesis_type": "phd", "thesis_year": "2007" }, { "id": "thesis:3157", "collection": "thesis", "collection_id": "3157", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-08172006-130145", "primary_object_url": { "basename": "main.pdf", "content": "final", "filesize": 2051167, "license": "other", "mime_type": "application/pdf", "url": "/3157/1/main.pdf", "version": "v3.0.0" }, "type": "thesis", "title": "Distributed Estimation and Control in Networked Systems", "author": [ { "family_name": "Gupta", "given_name": "Vijay", "clpid": "Gupta-Vijay" } ], "thesis_advisor": [ { "family_name": "Murray", "given_name": "Richard M.", "clpid": "Murray-R-M" }, { "family_name": "Hassibi", "given_name": "Babak", "clpid": "Hassibi-B" } ], "thesis_committee": [ { "family_name": "Murray", "given_name": "Richard M.", "clpid": "Murray-R-M" }, { "family_name": "Hassibi", "given_name": "Babak", "clpid": "Hassibi-B" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Schulman", "given_name": "Leonard J.", "clpid": "Schulman-L-J" }, { "family_name": "Vaidyanathan", "given_name": "P. P.", "clpid": "Vaidyanathan-P-P" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "Rapid advances in information processing, communication and sensing technologies have enabled more and more devices to be provided with embedded processors, networking capabilities and sensors. For the field of estimation and control, it is now possible to consider an architecture in which many simple components communicate and cooperate to achieve a joint team goal. This distributed (or networked) architecture promises much in terms of performance, reliability and simplicity of design; however, at the same time, it requires extending the traditional theories of control, communication and computation and, in fact, looking at a unified picture of the three fields. A systematic theory of how to design distributed systems is currently lacking.
\r\n\r\nThis dissertation takes the first steps towards understanding the effects of imperfect information flow in distributed systems from an estimation and control perspective and coming up with new design principles to counter these effects. Designing networked systems is difficult because such systems challenge two basic assumptions of traditional control theory - presence of a central node with access to all the information about the system and perfect transmission of information among components. We formulate and solve many problems that deal with the removal of one, or both, of these assumptions. The chief idea explored in this dissertation is the joint design of information flow and the control law. While traditional control design has concentrated on calculating the optimal control input by assuming a particular information flow between the components, our approach seeks to synthesize the optimal information flow along with the optimal control law that satisfies the constraints of the information flow. Thus besides the question of 'What should an agent do?', the questions of 'Whom should an agent talk to?', 'What should an agent communicate?', 'When should an agent communicate?' and so on also have to be answered. The design of the information flow represents an important degree of freedom available to the system designer that has hitherto largely been ignored. As we demonstrate in the dissertation, the joint design of information flow and the optimal control input satisfying the constraints of that information flow yields large improvements in performance over simply trying to fit traditional design theories on distributed systems.
\r\n\r\nWe begin by formulating a distributed control problem in which many agents in a formation need to cooperate to minimize a joint cost function. We provide numerical algorithms to synthesize the optimal constrained control law that involve solving linear equations only and hence are free from numerical issues plaguing the other approaches proposed in the literature. We then provide and analyze a model to understand the issue of designing the topology according to which the agents interact. The results are very surprising since there are cases when allowing communication to happen between two agents may, in fact, be detrimental to the performance.
\r\n\r\nWe then move on to consider the effects of communication channels on control performance. To counter such effects, we propose the idea of encoding information for the purpose of estimation and control prior to transmission. Although information theoretic techniques are not possible in our problem, we are able to solve for a recursive yet optimal encoder / decoder structure in many cases. This information flow design oriented approach has unique advantages such as being optimal for any packet drop pattern, being able to include the effect of known but random delays easily, letting us escape the limits set by reliability for transmission of data across a network by using intermediate nodes as 'repeaters' similar to a digital communication network and so on.
\r\n\r\nWe finally take a look at combining the effects of multiple sources of information and communication channels on estimation and control. We look at a distributed estimation problem in which, at every time step, only a subset out of many sensors can transmit information to the estimator. This is also a representative resource allocation problem. We propose the idea of stochastic communication patterns that allows us to include the effects of communication channels explicitly during system design. Thus, instead of tree-search based algorithms proposed in the literature, we provide stochastic scheduling algorithms that can take into account the random packet drop effect of the channels. We also consider a distributed control problem with switching topologies and solve for the optimal controller. The tools that we develop are applicable to many other scenarios and we demonstrate some of them in the dissertation.
\r\n\r\nAlong the way, we look at many other related problems in the dissertation. As an example, we provide initial results about the issue of robustness of a distributed system design to a malfunctioning agent. This notion is currently lacking in the control and estimation community, but has to be a part of any effective theory for designing networked or distributed systems.
", "doi": "10.7907/KWN2-X741", "publication_date": "2007", "thesis_type": "phd", "thesis_year": "2007" }, { "id": "thesis:86", "collection": "thesis", "collection_id": "86", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-01092007-152909", "primary_object_url": { "basename": "nieman_thesis.pdf", "content": "final", "filesize": 3673993, "license": "other", "mime_type": "application/pdf", "url": "/86/1/nieman_thesis.pdf", "version": "v2.0.0" }, "type": "thesis", "title": "Postdiction and the Effects of Spatial, Temporal, and Feature Compatibility on Sensory Integration", "author": [ { "family_name": "Nieman", "given_name": "Dylan Rhichard", "clpid": "Nieman-Dylan-Rhichard" } ], "thesis_advisor": [ { "family_name": "Shimojo", "given_name": "Shinsuke", "clpid": "Shimojo-S" } ], "thesis_committee": [ { "family_name": "Andersen", "given_name": "Richard A.", "clpid": "Andersen-R-A" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Allman", "given_name": "John Morgan", "clpid": "Allman-J-M" }, { "family_name": "Konishi", "given_name": "Masakazu", "clpid": "Konishi-M" }, { "family_name": "Shimojo", "given_name": "Shinsuke", "clpid": "Shimojo-S" } ], "local_group": [ { "literal": "div_biol" } ], "abstract": "The brain continually integrates stimuli over multiple sensory modalities and reconciles often disparate information into a unified, consistent representation of the surrounding environment. This process must be robust to differential neural latencies and imperfect alignments of spatial reference frames between sensory modalities. Numerous studies have examined the perception of multisensory stimuli with the presumption that multisensory integration is categorically different from within-modality integration. We looked at a variety of issues related to the updating of sensory reference frames and the integration of unimodal and multimodal stimuli over temporal and spatial disparities. Study 1 found simultaneous, opposite gaze-dependent aftereffects at the same retinal location for both depth and color, demonstrating the degree to which visual-coordinate space is gaze-contingent, not merely retinotopic. Study 2 found that the flash-lag effect, in which a flashed target is perceived as lagging behind a smoothly moving target, generalizes to third-order motion perception of cyclopean stimuli. Study 3 introduced a novel motion illusion which we termed the \"turn-point phantom,\" wherein the position of an abrupt orthogonal direction change is mislocalized backwards along the object's subsequent trajectory. This effect, like flash-lag, can only be adequately explained with postdiction. Study 4 explored the effect of passive head or body turns on spatial perception of visual and auditory stimuli and found systematic mislocalization of pre-turn stimuli in the direction of the turn. This mislocalization decayed with added delay between target and turn onset. Study 5 examined spatial and temporal disparity in visual-motor ventriloquism and found that early visual distracters were essentially equivalent, whereas the influence of late visual distracters diminished with increasing asynchrony. Study 6 found suppression of saccade latency induced by stimulus repetition in certain multisensory experimental contexts. Together, these studies provide numerous examples supporting the idea that sensory perception, both unimodal and multimodal, is postdictive in nature, involving integration of sensory information over a time window that includes, but does not end with, task-relevant stimulus presentation. Additionally, these results provide clues to the character and relevant parameters of the integration process.", "doi": "10.7907/QGYV-PR54", "publication_date": "2007", "thesis_type": "phd", "thesis_year": "2007" }, { "id": "thesis:1554", "collection": "thesis", "collection_id": "1554", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-04302007-164103", "primary_object_url": { "basename": "thesis.pdf", "content": "final", "filesize": 3123632, "license": "other", "mime_type": "application/pdf", "url": "/1554/1/thesis.pdf", "version": "v3.0.0" }, "type": "thesis", "title": "The Self-Replication and Evolution of DNA Crystals", "author": [ { "family_name": "Schulman", "given_name": "Rebecca Beth", "orcid": "0000-0003-4555-3162", "clpid": "Schulman-Rebecca-Beth" } ], "thesis_advisor": [ { "family_name": "Winfree", "given_name": "Erik", "clpid": "Winfree-E" } ], "thesis_committee": [ { "family_name": "Abu-Mostafa", "given_name": "Yaser S.", "clpid": "Abu-Mostafa-Y-S" }, { "family_name": "Yurke", "given_name": "Bernard", "clpid": "Yurke-B" }, { "family_name": "Winfree", "given_name": "Erik", "clpid": "Winfree-E" }, { "family_name": "Joyce", "given_name": "Gerald", "clpid": "Joyce-G" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Wang", "given_name": "Zhen-Gang", "clpid": "Wang-Zhen-Gang" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "How life began is still a mystery. While various theories suggest that life began in deep sea volcanic vents or that a world where life consisted predominantly of RNA molecules preceded us, there is no hard evidence to give shape to the chain of events that led to cellular life.
\r\n\r\nPerhaps the fundamental enigma of our origins is how life began to self-replicate in such a way that evolution could produce Earth's \"endless forms most beautiful.\" With the exception of biological organisms, we have no examples of self-replicating, evolving chemical systems, despite an extensive research program with the goal of identifying them.
\r\n\r\nIn this thesis, I construct a chemical system that is capable of the most basic self-replication and evolution. The system uses no enzymes or biological sequences, can support and replicate a combinatorial genome, and is completely autonomous. There are no fundamental obstacles to the replication by this system of much more complex sequences or to open-ended evolution.
\r\n\r\nThe design of the system is inspired by the work of Graham Cairns-Smith, who has proposed that life began with clay. Clays are tiny layered crystals; some clay crystals can contain one of several different patterns of atoms or molecules in each layer. The choice of patterns for the layers could be viewed as a sort of genome: it would be copied as the clay grew, and if the crystal broke, each new piece would inherit its pattern from the old piece and could replicate it in the same manner. If some patterns of layers could grow and reproduce faster than other patterns, crystals with faster-growing patterns would be selected for.
\r\n\r\nInstead of the atoms or small molecules of which clay consists, I use molecules consisting of 4-6 interwoven, synthetic DNA strands called DNA tiles to construct crystals that replicate and evolve as Cairns-Smith imagined. While the choice of construction material was influenced by ease of use -- in contrast to clay crystals, DNA tile crystals have been previously characterized and are easy to crystallize and image in the laboratory -- the choice was fundamentally made because DNA tile monomers are programmable, allowing us to create novel crystal morphologies rationally.
\r\n\r\nThe crystals I construct, termed \"zig-zag ribbons\", contain a sequence of information (\"a genome\") in each row. Growth of the ribbon adds rows, one at time, each of which contain an arrangement of DNA tiles that encode the same information sequence as the previous row. Altering the set of \"tiles\" used to assemble ribbons allows us to alter the alphabets for and the permitted lengths of sequences that can be copied.
\r\n\r\nI describe how to design tile sets that can replicate genomes with different alphabets and the kind of sequence evolution that is in theory possible with some simple tile sets. Altering the tile set can not only change the kinds of sequences that may be replicated, it can also make growth and splitting more robust. I show how to make changes to the crystals' design to prevent errors during growth and splitting and to reduce the rate of spontaneous generation of new crystals.
\r\n\r\nIt has been previously shown that DNA tile crystallization can be used to perform universal computation; I show that in theory crystals that can compute can undergo open-ended evolution as they try to produce more and more complex programs to take advantage of available growth resources. This mechanism is simple enough to potentially observe in the laboratory in the near future. In experiments, I demonstrate a much more basic kind of replication and evolution, in which zig-zag ribbons maintain a preference for a certain width into a second generation.
\r\n\r\nThis work suggests that the concept of a self-replicating chemistry is closely related to the concept of a chemistry that can store information and compute. It is only by clearly understanding how chemistry can perform these latter tasks that we can hope to understand how self-replication and evolution can occur, and by implication, understand how life might have begun.
", "doi": "10.7907/3F8C-9D50", "publication_date": "2007", "thesis_type": "phd", "thesis_year": "2007" }, { "id": "thesis:2404", "collection": "thesis", "collection_id": "2404", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-06022006-154329", "primary_object_url": { "basename": "drummond-thesis.pdf", "content": "final", "filesize": 2942072, "license": "other", "mime_type": "application/pdf", "url": "/2404/1/drummond-thesis.pdf", "version": "v2.0.0" }, "type": "thesis", "title": "Misfolding Dominates Protein Evolution", "author": [ { "family_name": "Drummond", "given_name": "David Allan", "orcid": "0000-0001-7018-7059", "clpid": "Drummond-David-Allan" } ], "thesis_advisor": [ { "family_name": "Arnold", "given_name": "Frances Hamilton", "orcid": "0000-0002-4027-364X", "clpid": "Arnold-F-H" } ], "thesis_committee": [ { "family_name": "Arnold", "given_name": "Frances Hamilton", "orcid": "0000-0002-4027-364X", "clpid": "Arnold-F-H" }, { "family_name": "Adami", "given_name": "Christoph Carl", "orcid": "0000-0002-2915-9504", "clpid": "Adami-C-C" }, { "family_name": "Winfree", "given_name": "Erik", "orcid": "0000-0002-5899-7523", "clpid": "Winfree-E" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Elowitz", "given_name": "Michael B.", "orcid": "0000-0002-1221-0967", "clpid": "Elowitz-M-B" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "The diverse array of protein functions depends upon these molecules' reliable ability to fold into the native structures determined by their amino-acid sequences. Because mutations that alter a protein's sequence frequently disrupt its folding, protein evolution explores protein sequence space conservatively, either by point mutations or recombination between related sequences. Attempts to engineer proteins by co-opting the evolutionary algorithm have also largely proceeded by the stepwise accumulation of beneficial mutations. Other strategies for directed evolution have focused on introducing many mutations at once as a way to increase the likelihood of finding improved variants, attempting to balance higher mutational diversity with lower retention of folding. Using simple models, I explore this tradeoff and find that protein misfolding dominates whether increasing mutation levels increase the number of improved variants. I analyze results of a popular mutagenesis protocol, error-prone PCR, for evidence that coupling between mutations might favor higher mutation levels, as claimed by several groups. A comparison of high-mutation-rate mutagenesis to protein recombination between distantly related proteins reveals qualitative differences in protein tolerance for sequence changes introduced by each method. Mutational tolerance may also be reflected in the rate at which proteins accumulate sequence changes over evolutionary time; why proteins evolve at different rates remains a major open question in biology. An analysis of rate determinants suggests that one major variable, linked to how highly expressed the encoding gene is, dominates the rate of yeast protein evolution. To explain this trend, I hypothesize that proteins are selected to fold properly despite mistranslation, a property I call translational robustness, and test it using genomic data. To examine protein evolution at a higher level of detail, a large-scale simulation is constructed in which simulated organisms, with genomes containing genes expressing computationally foldable proteins at different levels, evolve over millions of generations with protein misfolding imposing the only fitness cost. The results suggest that protein misfolding suffices to explain many significant trends in genome evolution observed across taxa, predict a novel genomic trend which is then identified in yeast, and create insight into the causes of evolutionary rate variation in proteins.", "doi": "10.7907/DH8E-2N10", "publication_date": "2006", "thesis_type": "phd", "thesis_year": "2006" }, { "id": "thesis:3062", "collection": "thesis", "collection_id": "3062", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-08092005-100809", "primary_object_url": { "basename": "Final_Thesis.pdf", "content": "final", "filesize": 7751861, "license": "other", "mime_type": "application/pdf", "url": "/3062/1/Final_Thesis.pdf", "version": "v2.0.0" }, "type": "thesis", "title": "Signal Integrity Issues in High-Speed Wireline Links: Analysis and Integrated System Solutions", "author": [ { "family_name": "Analui", "given_name": "Behnam", "clpid": "Analui-Behnam" } ], "thesis_advisor": [ { "family_name": "Hajimiri", "given_name": "Ali", "clpid": "Hajimiri-A" } ], "thesis_committee": [ { "family_name": "Hajimiri", "given_name": "Ali", "clpid": "Hajimiri-A" }, { "family_name": "Rutledge", "given_name": "David B.", "clpid": "Rutledge-D-B" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Weinreb", "given_name": "Sander", "clpid": "Weinreb-S" }, { "family_name": "Tai", "given_name": "Yu-Chong", "clpid": "Tai-Yu-Chong" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "This work focuses on the basic signal integrity issues of high-speed wireline links. It bridges the gap between optimum system design and circuit design for such links by: (1) understanding the effects of the system parameters on the bit error rate (BER), (2) introducing circuit architectures for the realization of systems that minimize the BER, and (3) demonstrating integrated circuit prototypes that verify the solutions.
\r\n\r\nFirst, we develop a theory that analytically relates the data link BER to the system characteristics, e.g., the channel response, the pre-amplifier bandwidth, and the transmitter clock jitter. We generate the BER contours to find the optimum receiver bandwidth as well as the optimum sampling point and its associated timing margin. We also develop the theory of the data-dependent jitter (DDJ), which is a significant component of the timing jitter in high-speed links. We provide an analytical distribution function for the DDJ of an arbitrary linear time-invariant system and include the impact of the DDJ on the BER.
\r\n\r\nSecond, we propose a bandwidth enhancement method for wideband amplifiers. This is useful for the realization of high-speed links in technologies that suffer from large parasitic components. The method leverages two-port broadband matching to enable amplifier stages to achieve their maximum gain-bandwidth product. We demonstrate a 10Gb/s CMOS 0.18um amplifier with this technique that has 2.4 times the bandwidth improvement over a design that does not apply the technique.
\r\n\r\nThird, we develop an eye-opening monitor (EOM) that enables full integration of adaptive equalizers. The EOM evaluates the signal eye diagram quality and reports a quantitative measure, which is correlated to the signal integrity. We demonstrate a prototype in 0.13um standard CMOS that operates up to 12.5Gb/s and has 68dB error dynamic range.
\r\n\r\nFinally, we introduce an instantaneous clockless demultiplexer for burst-mode communication applications. We propose a clockless finite state machine that recovers and demultiplexes the received burst of data instantaneously. The architecture consists of a combinational logic structure and a bit-period-delayed feedback loop. We demonstrate a 1:2 clockless demultiplexer based on this concept in SiGe BiCMOS technology that operates at 7.5Gb/s.
", "doi": "10.7907/GGMB-0J23", "publication_date": "2006", "thesis_type": "phd", "thesis_year": "2006" }, { "id": "thesis:2846", "collection": "thesis", "collection_id": "2846", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-07102006-114157", "primary_object_url": { "basename": "01Neil_Introduction.pdf", "content": "final", "filesize": 102579, "license": "other", "mime_type": "application/pdf", "url": "/2846/1/01Neil_Introduction.pdf", "version": "v2.0.0" }, "type": "thesis", "title": "Development of Audiovisual Integration in Human Infants: The Effects of Spatial and Temporal Congruency and Incongruency on Response Latencies", "author": [ { "family_name": "Neil", "given_name": "Patricia Ann", "clpid": "Neil-Patricia-Ann" } ], "thesis_advisor": [ { "family_name": "Shimojo", "given_name": "Shinsuke", "clpid": "Shimojo-S" } ], "thesis_committee": [ { "family_name": "Adolphs", "given_name": "Ralph", "clpid": "Adolphs-R" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Perona", "given_name": "Pietro", "clpid": "Perona-P" }, { "family_name": "Andersen", "given_name": "Richard A.", "clpid": "Andersen-R-A" }, { "family_name": "Shimojo", "given_name": "Shinsuke", "clpid": "Shimojo-S" } ], "local_group": [ { "literal": "div_biol" } ], "abstract": "Every day we are inundated with a mass of sensory inputs providing a continual stream of relevant and irrelevant, redundant and conflicting, information about the external world. Mature brains are very capable in integrating this confusion of input into a unified percept, but this is a non-trivial task for infants, whose brains and sensory systems are still immature at birth and who rely on their current level of integration and interaction of these inputs in order to shape their future development. Failure in being able to properly process basic sensory interactions has been implicated in higher-level developmental problems like attentional or autistic spectrum disorders. Numerous studies have looked at how adults perceive and react to multisensory stimuli, including findings of improved response latencies and target detection for spatially and temporally congruent stimuli, but much less is known about the development of multisensory integration or how spatial or temporal disparities effect sensory interactions in young babies. We examined the role of spatial and temporal congruency and incongruency on the response latencies of infants under ten months of age orienting toward an audiovisual stimulus at +/-25 degrees and/or +/-45 degrees. In Study 1, we found the beginnings of adult-style non-linear integration for spatially and temporally congruent audiovisual targets in 8\u201310 month olds, but not in younger infants, as well as indications of a differential developmental profile for binaural versus monaural processing. In Studies 2 and 3, spatial and temporal disparities were found to significantly lengthen infants\u2019 response latencies to an audiovisual target. We also found clear indications of developmental changes for all three spatial and temporal conditions, as well as key dependencies in relative position, temporal order, and sensory dominance.", "doi": "10.7907/XPGR-QT95", "publication_date": "2006", "thesis_type": "phd", "thesis_year": "2006" }, { "id": "thesis:2011", "collection": "thesis", "collection_id": "2011", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-05242006-170918", "primary_object_url": { "basename": "thesis.pdf", "content": "final", "filesize": 1812990, "license": "other", "mime_type": "application/pdf", "url": "/2011/1/thesis.pdf", "version": "v4.0.0" }, "type": "thesis", "title": "Heterogeneous Congestion Control Protocols", "author": [ { "family_name": "Tang", "given_name": "Ao (Kevin)", "clpid": "Tang-Ao-Kevin" } ], "thesis_advisor": [ { "family_name": "Low", "given_name": "Steven H.", "clpid": "Low-S-H" } ], "thesis_committee": [ { "family_name": "Low", "given_name": "Steven H.", "clpid": "Low-S-H" }, { "family_name": "Hassibi", "given_name": "Babak", "clpid": "Hassibi-B" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Doyle", "given_name": "John Comstock", "clpid": "Doyle-J-C" }, { "family_name": "Chandy", "given_name": "K. Mani", "clpid": "Chandy-K-M" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "Homogeneity of price is an implicit yet fundamental assumption underlying price based resource allocation theory. In this thesis, we study the effects of relaxing this assumption by examining a concrete engineering system (network with heterogeneous congestion control protocols). The behavior of the system turns out to be very different from the homogeneous case and can potentially be much more complicated. A systematic theory is developed that includes all major properties of equilibrium of the system such as existence, uniqueness, optimality, and stability. In addition to analysis, we also present numerical examples, simulations, and experiments to illustrate the theory and verify its predictions.
\r\n\r\nWhen heterogeneous congestion control protocols that react to different pricing signals share the same network, the resulting equilibrium can no longer be interpreted as a solution to the standard utility maximization problem as the current theory suggests. After introducing a mathematical formulation of network equilibrium for multi-protocol networks, we prove the existence of equilibrium under mild assumptions. For almost all networks, the equilibria are locally unique. They are finite and odd in number. They cannot all be locally stable unless the equilibrium is globally unique. We also derive two conditions for global uniqueness. By identifying an optimization problem associated with every equilibrium, we show that every equilibrium is Pareto efficient and provide an upper bound on efficiency loss due to pricing heterogeneity. Both intra-protocol and inter-protocol fairness are then discussed. On dynamics, various stability results are provided. In particular it is shown that if the degree of pricing heterogeneity is small enough, the network equilibrium is not only unique but also locally stable. Finally, a distributed algorithm is proposed to steer a network to the unique equilibrium that maximizes the aggregate utility, by only updating a linear parameter in the sources' algorithms in a slow timescale.
", "doi": "10.7907/eh43-pa83", "publication_date": "2006", "thesis_type": "phd", "thesis_year": "2006" }, { "id": "thesis:2424", "collection": "thesis", "collection_id": "2424", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-06032005-140944", "primary_object_url": { "basename": "Thesis.pdf", "content": "final", "filesize": 1268526, "license": "other", "mime_type": "application/pdf", "url": "/2424/1/Thesis.pdf", "version": "v3.0.0" }, "type": "thesis", "title": "Networks of Relations", "author": [ { "family_name": "Cook", "given_name": "Matthew M.", "clpid": "Cook-Matthew-M" } ], "thesis_advisor": [ { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" } ], "thesis_committee": [ { "family_name": "Winfree", "given_name": "Erik", "clpid": "Winfree-E" }, { "family_name": "Umans", "given_name": "Christopher M.", "clpid": "Umans-C-M" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Schulman", "given_name": "Leonard J.", "clpid": "Schulman-L-J" }, { "family_name": "Abu-Mostafa", "given_name": "Yaser S.", "clpid": "Abu-Mostafa-Y-S" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "Relations are everywhere. In particular, we think and reason in terms of mathematical and English sentences that state relations. However, we teach our students much more about how to manipulate functions than about how to manipulate relations. Consider functions. We know how to combine functions to make new functions, how to evaluate functions efficiently, and how to think about compositions of functions. Especially in the area of boolean functions, we have become experts in the theory and art of designing combinations of functions to yield what we want, and this expertise has led to techniques that enable us to implement mind-bogglingly large yet efficient networks of such functions in hardware to help us with calculations. If we are to make progress in getting machines to be able to reason as well as they can calculate, we need to similarly develop our understanding of relations, especially their composition, so we can develop techniques to help us bridge between the large and small scales. There has been some important work in this area, ranging from practical applications such as relational databases to extremely theoretical work in universal algebra, and sometimes theory and practice manage to meet, such as in the programming language Prolog, or in the probabilistic reasoning methods of artificial intelligence. However, the real adventure is yet to come, as we learn to develop a better understanding of how relations can efficiently and reliably be composed to get from a low level representation to a high level representation, as this understanding will then allow the development of automated techniques to do this on a grand scale, finally enabling us to build machines that can reason as amazingly as our contemporary machines can calculate.
\r\n\r\nThis thesis explores new ground regarding the composition of relations into larger relational structures. First of all a foundation is laid by examining how networks of relations might be used for automated reasoning. We define exclusion networks, which have close connections with the areas of constraint satisfaction problems, belief propagation, and even boolean circuits. The foundation is laid somewhat deeper than usual, taking us inside the relations and inside the variables to see what is the simplest underlying structure that can satisfactorily represent the relationships contained in a relational network. This leads us to define zipper networks, an extremely low-level view in which the names of variables or even their values are no longer necessary, and relations and variables share a common substrate that does not distinguish between the two. A set of simple equivalence operations is found that allows one to transform a zipper network while retaining its solution structure, enabling a relation-variable duality as well as a canonical form on linear segments. Similarly simple operations allow automated deduction to take place, and these operations are simple and uniform enough that they are easy to imagine being implemented by biological neural structures.
\r\n\r\nThe canonical form for linear segments can be represented as a matrix, leading us to matrix networks. We study the question of how we can perform a change of basis in matrix networks, which brings us to a new understanding of Valiant's recent holographic algorithms, a new source of polynomial time algorithms for counting problems on graphs that would otherwise appear to take exponential time. We show how the holographic transformation can be understood as a collection of changes of basis on individual edges of the graph, thus providing a new level of freedom to the method, as each edge may now independently choose a basis so as to transform the matrices into the required form.
\r\n\r\nConsideration of zipper networks makes it clear that \"fan-out,\" i.e., the ability to duplicate information (for example allowing a variable to be used in many places), is most naturally itself represented as a relation along with everything else. This is a notable departure from the traditional lack of representation for this ability. This deconstruction of fan-out provides a more general model for combining relations than was provided by previous models, since we can examine both the traditional case where fan-out (the equality relation on three variables) is available and the more interesting case where its availability is sub ject to the same limitations as the availability of other relations. As we investigate the composition of relations in this model where fan-out is explicit, what we find is very different from what has been found in the past.
\r\n\r\nFirst of all we examine the relative expressive power of small relations: For each relation on three boolean variables, we examine which others can be implemented by networks built solely from that relation. (We also find, in each of these cases, the complexity of deciding whether such a network has a solution. We find that solutions can be found in polynomial time for all but one case, which is NP-complete.) For the question of which relations are able to implement which others, we provide an extensive and complete answer in the form of a hierarchy of relative expressive power for these relations. The hierarchy for relations is more complex than Post's well-known comparable hierarchy for functions, and parts of it are particularly difficult to prove. We find an explanation for this phenomenon by showing that in fact, the question of whether one relation can implement another (and thus should be located above it in the hierarchy) is undecidable. We show this by means of a complicated reduction from the halting problem for register machines. The hierarchy itself has a lot of structure, as it is rarely the case that two ternary boolean relations are equivalent. Often they are comparable, and often they are incomparable\u2014the hierarchy has quite a bit of width as well as depth. Notably, the fan-out relation is particularly difficult to implement; only a very few relations are capable of implementing it. This provides an additional ex post facto justification for considering the case where fan-out is absent: If you are not explicitly provided with fan-out, you are unlikely to be able to implement it.
\r\n\r\nThe undecidability of the hierarchy contrasts strongly with the traditional case, where the ubiquitous availability of fan-out causes all implementability questions to collapse into a finite decidable form. Thus we see that for implementability among relations, fan-out leads to undecidability. We then go on to examine whether this result might be taken back to the world of functions to find a similar difference there. As we study the implementability question among functions without fan-out, we are led directly to questions that are independently compelling, as our functional implementability question turns out to be equivalent to asking what can be computed by sets of chemical reactions acting on a finite number of species. In addition to these chemical reaction networks, several other nondeterministic systems are also found to be equivalent in this way to the implementability question, namely, Petri nets, unordered Fractran, vector addition systems, and \"broken\" register machines (whose decrement instruction may fail even on positive registers). We prove equivalences between these systems.
\r\n\r\nWe find several interesting results in particular for chemical reaction networks, where the standard model has reaction rates that depend on concentration. In this setting, we analyze questions of possibility as well as questions of probability. The question of the possibility of reaching a target state turns out to be equivalent to the reachability question for Petri nets and vector addition systems, which has been well studied. We provide a new proof that a form of this reachability question can be decided by primitive recursive functions. Ours is the first direct proof of this relationship, avoiding the traditional excursion to Diophantine equations, and thus providing a crisper picture of the relationship between Karp's coverability tree and primitive recursive functions.
\r\n\r\nIn contrast, the question of finding the probability (according to standard chemical kinetics) of reaching a given target state turns out to be undecidable. Another way of saying this is that if we wish to distinguish states with zero probability of occurring from states with positive probability of occurring, we can do so, but if we wish to distinguish low probability states from high probability states, there is no general way to do so. Thus, if we wish to use a chemical reaction network to perform a computation, then if we insist that the network must always get the right answer, we will only be able to use networks with limited computational power, but if we allow just the slightest probability of error, then we can use networks with Turing-universal computational ability. This power of probability is quite surprising, especially when contrasted with the conventional computational complexity belief that BPP = P.
\r\n\r\nExploring the source of this probabilistic power, we find that the probabilities guiding the network need to depend on the concentrations (or perhaps on time)\u2014fixed probabilities aren\u2019t enough on their own to achieve this power. In the language of Petri nets, if one first picks a transition at random, and then fires it if it is enabled, then the probability of reaching a particular target state can be calculated to arbitrary precision, but if one first picks a token at random, and then fires an enabled transition that will absorb that token, then the probability of reaching a particular target state cannot in general be calculated to any precision whatsoever.
\r\n\r\nIn short, what started as a simple thorough exploration of the power of composition of relations has led to many decidability and complexity questions that at first appear completely unrelated, but turn out to combine to paint a coherent picture of the relationship between relations and functions, implementability and reachability, possibility and probability, and decidability and undecidability.
", "doi": "10.7907/CVKM-D684", "publication_date": "2005", "thesis_type": "phd", "thesis_year": "2005" }, { "id": "thesis:2002", "collection": "thesis", "collection_id": "2002", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-05242005-162056", "primary_object_url": { "basename": "savarese_thesis.pdf", "content": "final", "filesize": 6963221, "license": "other", "mime_type": "application/pdf", "url": "/2002/1/savarese_thesis.pdf", "version": "v2.0.0" }, "type": "thesis", "title": "Shape Reconstruction from Shadows and Reflections", "author": [ { "family_name": "Savarese", "given_name": "Silvio", "clpid": "Savarese-Silvio" } ], "thesis_advisor": [ { "family_name": "Perona", "given_name": "Pietro", "clpid": "Perona-P" } ], "thesis_committee": [ { "family_name": "Perona", "given_name": "Pietro", "clpid": "Perona-P" }, { "family_name": "Psaltis", "given_name": "Demetri", "clpid": "Psaltis-D" }, { "family_name": "Arvo", "given_name": "James R.", "clpid": "Arvo-J-R" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Shimojo", "given_name": "Shinsuke", "clpid": "Shimojo-S" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "Measuring automatically the shape of physical objects in order to obtain corresponding digital models has become a useful, often indispensable, tool in design, engineering, art conservation, computer graphics, medicine and science. Machine vision has proven to be more appealing than competing technologies. Ideally, we would like to be able to acquire digital models of generic objects by simply walking around the scene, while filming with a handheld camcorder. Thus, one of the main challenges in modern machine vision is to develop algorithms that: i) are inexpensive, fast and accurate; ii) can handle objects with arbitrary appearance properties and shape; and iii) need little or no user intervention.
\r\n\r\nIn this thesis, we address both issues. In the first part, we present a novel 3D reconstruction technique which makes use of minimal and inexpensive equipment. We call this technique \"shadow carving\". We explore the information contained in the shadows that an object casts upon itself. An algorithm is provided that makes use of this information. The algorithm iteratively recovers an estimate of the object which i) approximates the object\u2019s shape more and more closely; and ii) is provably an upper bound to the object's shape. Shadow carving is the first technique to incorporate \"shadow\" information in a multi-view shape recovery framework. We have implemented our approach in a simple table-top system and validated our algorithm by recovering the shape of real objects.
\r\n\r\nIt is well known that vision-based 3D scanning systems handle specular or highly reflective surfaces only poorly. The cause of this deficiency is most likely not intrinsic, but rather due to our lack of understanding of the relevant cues. In the second part of this thesis, we focus on how to promote mirror reflections from \"noise\" to \"signal\". We first present a geometrical and algebraic characterization of how a patch of the scene is mapped into an image by a mirror surface of given shape. We then develop solutions to the inverse problem of deriving surface shape from mirror reflections in a single image. We validate our theoretical results with both numerical simulations and experiments with real surfaces.
\r\n\r\nA third goal of this thesis is advancing our understanding of human perception of shape from reflections. Although the idea of perception of shape from different visual cues (e.g., shading, texture, etc.) has been extensively discussed in the past, little is known to what extent highlights and specular reflections carry useful information for shape perception. We use psychophysics to study this capability. Our goal is to provide a benchmark, as well as inspire possible technical approaches, for our computational work. We find that surprisingly, humans are very poor at judging the shape of mirror surfaces when additional visual cues (i.e., contour, shading, stereo, texture) are not visible.
", "doi": "10.7907/FH0V-3M10", "publication_date": "2005", "thesis_type": "phd", "thesis_year": "2005" }, { "id": "thesis:2818", "collection": "thesis", "collection_id": "2818", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-07072004-154316", "primary_object_url": { "basename": "thesis.pdf", "content": "final", "filesize": 2283527, "license": "other", "mime_type": "application/pdf", "url": "/2818/1/thesis.pdf", "version": "v3.0.0" }, "type": "thesis", "title": "Interval Modulation: A New Paradigm for the Design of High Speed Communication Systems", "author": [ { "family_name": "Mukhtar", "given_name": "Saleem", "clpid": "Mukhtar-Saleem" } ], "thesis_advisor": [ { "family_name": "Hajimiri", "given_name": "Ali", "orcid": "0000-0001-6736-8019", "clpid": "Hajimiri-A" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Sternberg", "given_name": "Paul W.", "orcid": "0000-0002-7699-0173", "clpid": "Sternberg-P-W" } ], "thesis_committee": [ { "family_name": "Koch", "given_name": "Christof", "orcid": "0000-0001-6482-8067", "clpid": "Koch-C" }, { "family_name": "Hajimiri", "given_name": "Ali", "orcid": "0000-0001-6736-8019", "clpid": "Hajimiri-A" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" }, { "family_name": "Low", "given_name": "Steven H.", "orcid": "0000-0001-6476-3048", "clpid": "Low-S-H" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "In this thesis we propose a new, biologically inspired, paradigm for the design of high speed communication systems. The paradigm consists of a new modulation format referred to as Interval Modulation (IM). In order to transmit data in an efficient manner using this format, new coding techniques are needed. In this thesis we propose a coding technique based on variable length to variable length prefix trees and code construction algorithms are outlined. These codes are referred to as Interval Modulation Codes (IMC). Furthermore, data encoded with this modulation format cannot be transmitted or received using conventional synchronous CDR based receivers. In this thesis we outline a new asynchronous circuit architecture for both the transmitter and receiver. The architecture is based on active delay lines and eliminates the need for clock recovery.", "doi": "10.7907/6F03-MP11", "publication_date": "2005", "thesis_type": "phd", "thesis_year": "2005" }, { "id": "thesis:2137", "collection": "thesis", "collection_id": "2137", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-05272004-163315", "primary_object_url": { "basename": "phd.pdf", "content": "final", "filesize": 1071633, "license": "other", "mime_type": "application/pdf", "url": "/2137/1/phd.pdf", "version": "v3.0.0" }, "type": "thesis", "title": "Optimized Network Data Storage and Topology Control", "author": [ { "family_name": "Jiang", "given_name": "Anxiao (Andrew)", "orcid": "0000-0002-0120-7930", "clpid": "Jiang-Anxiao-Andrew" } ], "thesis_advisor": [ { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" } ], "thesis_committee": [ { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Chandy", "given_name": "K. Mani", "clpid": "Chandy-K-M" }, { "family_name": "Blaum", "given_name": "Mario", "clpid": "Blaum-M" }, { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" }, { "family_name": "Low", "given_name": "Steven H.", "clpid": "Low-S-H" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "This thesis addresses two key challenges for network data-storage systems: optimizing data placement for highly efficient and robust data access, and constructing network topologies that facilitate data transmission scalable to both network sizes and network dynamics. It focuses on two new topics \u2014 data placement using erasure-correcting codes, and topology control for nodes in normed spaces. The first topic generalizes traditional file-assignment problems, and has the distinct feature of interleavingly placing data in networks. The second topic emphasizes the construction of network topologies that achieve excellent global performance in comprehensive measurements, through purely local decisions on connectivity. The results of the thesis deepen the current understanding on these important and intriguing topics, and follow a mathematically rigorous approach.
\r\n", "doi": "10.7907/91R7-MH71", "publication_date": "2004", "thesis_type": "phd", "thesis_year": "2004" }, { "id": "thesis:2199", "collection": "thesis", "collection_id": "2199", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-05282004-170744", "primary_object_url": { "basename": "thesis.pdf", "content": "final", "filesize": 1024560, "license": "other", "mime_type": "application/pdf", "url": "/2199/1/thesis.pdf", "version": "v3.0.0" }, "type": "thesis", "title": "On Source Coding for Networks", "author": [ { "family_name": "Fleming", "given_name": "Michael Ian James", "clpid": "Fleming-Michael-Ian-James" } ], "thesis_advisor": [ { "family_name": "Effros", "given_name": "Michelle", "clpid": "Effros-M" } ], "thesis_committee": [ { "family_name": "Effros", "given_name": "Michelle", "clpid": "Effros-M" }, { "family_name": "Hassibi", "given_name": "Babak", "clpid": "Hassibi-B" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" }, { "family_name": "Low", "given_name": "Steven H.", "clpid": "Low-S-H" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "In this thesis, I examine both applied and theoretical issues in network source coding.
\r\n\r\nThe applied results focus on the construction of locally rate-distortion-optimal vector quantizers for networks. I extend an existing vector quantizer design algorithm for arbitrary network topologies [1] to allow for the use of side information at the decoder and for the presence of channel errors. I show how to implement the algorithm and use it to design codes for several different systems. The implementation treats both fixed-rate and variable-rate quantizer design and includes a discussion of convergence and complexity. Experimental results for several different systems demonstrate in practice some of the potential performance benefits (in terms of rate, distortion, and functionality) of incorporating a network's topology into the design of its data compression system.
\r\n\r\nThe theoretical work covers several topics. Firstly, for a system with some side information known at both the encoder and the decoder, and some known only at the decoder, I derive the rate-distortion function and evaluate it for binary symmetric and Gaussian sources. I then apply the results for binary sources in evaluating the binary symmetric rate-distortion function for a system where the presence of side information at the decoder is unreliable. Previously, only upper and lower bounds were known for that problem. Secondly, I address with an example the question of whether feedback from a decoder to an encoder ever enlarges the achievable rate region for lossless network source coding of memoryless sources. Thirdly, I show how cutset methods can yield quick and simple rate-distortion converses for any source coding network. Finally, I present rate-distortion results for two different broadcast source coding systems.
", "doi": "10.7907/CY48-QJ71", "publication_date": "2004", "thesis_type": "phd", "thesis_year": "2004" }, { "id": "thesis:1591", "collection": "thesis", "collection_id": "1591", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-05032004-153842", "primary_object_url": { "basename": "marc.riedel.phd.pdf", "content": "final", "filesize": 2195990, "license": "other", "mime_type": "application/pdf", "url": "/1591/1/marc.riedel.phd.pdf", "version": "v2.0.0" }, "type": "thesis", "title": "Cyclic Combinational Circuits", "author": [ { "family_name": "Riedel", "given_name": "Marcus D.", "orcid": "0000-0002-3318-346X", "clpid": "Riedel-Marcus-D" } ], "thesis_advisor": [ { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" } ], "thesis_committee": [ { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Martin", "given_name": "Alain J.", "clpid": "Martin-A-J" }, { "family_name": "Hajimiri", "given_name": "Ali", "clpid": "Hajimiri-A" }, { "family_name": "Viterbi", "given_name": "Andrew", "clpid": "Viterbi-A" }, { "family_name": "Winfree", "given_name": "Erik", "clpid": "Winfree-E" }, { "family_name": "Abu-Mostafa", "given_name": "Yaser S.", "clpid": "Abu-Mostafa-Y-S" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "A collection of logic gates forms a combinational circuit if the outputs can be described as Boolean functions of the current input values only. Optimizing combinational circuitry, for instance, by reducing the number of gates (the area) or by reducing the length of the signal paths (the delay), is an overriding concern in the design of digital integrated circuits.
\r\n\r\nThe accepted wisdom is that combinational circuits must have acyclic (i.e., loop-free or feed-forward) topologies. In fact, the idea that \"combinational\" and \"acyclic\" are synonymous terms is so thoroughly ingrained that many textbooks provide the latter as a definition of the former. And yet simple examples suggest that this is incorrect. In this dissertation, we advocate the design of cyclic combinational circuits (i.e., circuits with loops or feedback paths). We demonstrate that circuits can be optimized effectively for area and for delay by introducing cycles.
\r\n\r\nOn the theoretical front, we discuss lower bounds and we show that certain cyclic circuits are one-half the size of the best possible equivalent acyclic implementations. On the practical front, we describe an efficient approach for analyzing cyclic circuits, and we provide a general framework for synthesizing such circuits. On trials with industry-accepted benchmark circuits, we obtained significant improvements in area and delay in nearly all cases. Based on these results, we suggest that it is time to re-write the definition: combinational might well mean cyclic.
", "doi": "10.7907/410B-XR25", "publication_date": "2004", "thesis_type": "phd", "thesis_year": "2004" }, { "id": "thesis:2325", "collection": "thesis", "collection_id": "2325", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-05312004-205111", "primary_object_url": { "basename": "mythesis.pdf", "content": "final", "filesize": 602531, "license": "other", "mime_type": "application/pdf", "url": "/2325/1/mythesis.pdf", "version": "v3.0.0" }, "type": "thesis", "title": "Data Collection and Distribution in Sensory Networks", "author": [ { "family_name": "Florens", "given_name": "C\u00e9dric Jean Paul", "clpid": "Florens-C\u00e9dric-Jean-Paul" } ], "thesis_advisor": [ { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" } ], "thesis_committee": [ { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" }, { "family_name": "Hassibi", "given_name": "Babak", "clpid": "Hassibi-B" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Estabrook", "given_name": "Polly", "clpid": "Estabrook-P" }, { "family_name": "Low", "given_name": "Steven H.", "clpid": "Low-S-H" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "The deployment of large-scale, low-cost, low-power, multifunctional sensory networks brings forward numerous and diverse research challenges. Critical to the design of systems that must operate under extreme resource constraints, the understanding of the fundamental performance limits of sensory networks is a research topic of particular importance. This thesis examines, in this respect, an essential function of sensory networks, viz., data collection, that is, the aggregation at the user location of information gathered by sensor nodes.
\r\n\r\nIn the first part of this dissertation we study, via simple discrete mathematical models, the time performance of the data collection and data distribution tasks in sensory networks. Specifically, we derive the minimum delay in collecting sensor data for networks of various topologies such as line, multi-line, tree and give corresponding optimal scheduling strategies assuming that the amount of data observed at each node is finite and known at the beginning of the data collection phase. Furthermore, we bound the data collection time on general graph networks.
\r\n\r\nIn the second part of this dissertation we take the view that the amount of data collected at a node is random and study the statistics of the data collection time. Specifically, we analyze the average minimum delay in collecting randomly located/distributed sensor data for networks of various topologies when the number of nodes becomes large. Furthermore, we analyze the impact of various parameters such as lack of synchronization, size of packet, transmission range, and channel packet erasure probability on the optimal time performance. Our analysis applies to directional antenna systems as well as omnidirectional ones. We conclude our study with a simple comparative analysis showing the respective advantages of the two systems.
\r\n", "doi": "10.7907/ZK3J-VB92", "publication_date": "2004", "thesis_type": "phd", "thesis_year": "2004" }, { "id": "thesis:1800", "collection": "thesis", "collection_id": "1800", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-05142004-214312", "primary_object_url": { "basename": "tkacenko_phd_thesis.pdf", "content": "final", "filesize": 1272977, "license": "other", "mime_type": "application/pdf", "url": "/1800/1/tkacenko_phd_thesis.pdf", "version": "v2.0.0" }, "type": "thesis", "title": "Optimization Algorithms for Realizable Signal-Adapted Filter Banks", "author": [ { "family_name": "Tkacenko", "given_name": "Andre", "clpid": "Tkacenko-Andre" } ], "thesis_advisor": [ { "family_name": "Vaidyanathan", "given_name": "P. P.", "clpid": "Vaidyanathan-P-P" } ], "thesis_committee": [ { "family_name": "Vaidyanathan", "given_name": "P. P.", "clpid": "Vaidyanathan-P-P" }, { "family_name": "Hassibi", "given_name": "Babak", "clpid": "Hassibi-B" }, { "family_name": "Vrcelj", "given_name": "Bojan", "clpid": "Vrcelj-B" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "Multirate filter banks are fundamental systems commonly used in digital signal processing (DSP). Typically, they are used to decompose a discrete-time signal into a set of frequency selective components called subband signals. Filter banks have been found to be useful for lossy data compression schemes such as MP3 and JPEG 2000, denoising, and signal estimation. In the last decade, transmultiplexers, the dual structures of multirate filter banks, have been shown to be useful in digital communications systems such as discrete multitone (DMT) systems for channel equalization and inter/intra-symbol interference cancellation in the presence of noise.
\r\n\r\nRecently, a special type of filter bank adapted to its input known as the principal component filter bank (PCFB) has been shown to be simultaneously optimal for a wide variety of objectives. Such filter banks are not only optimal for relevant data compression type objectives such as coding gain and multiresolution, but also for digital communications type objectives such as power minimization, when the filter bank is implemented in its transmultiplexer form. The only problem is that PCFBs, which are defined over classes of paraunitary (PU) filter banks, are only known to exist for certain classes. In particular, PCFBs are in general known to exist only in the extremal cases where the analysis/synthesis polyphase matrix has zero memory and doubly infinite memory, respectively. Furthermore, for many practical cases of inputs, the filters corresponding to the infinite-order PCFB have ideal bandpass response and are as such unrealizable. When the polyphase matrix has finite memory or a finite impulse response (FIR), it is believed that PCFBs do not exist, although this has not yet been formally proven in the literature.
\r\n\r\nThe main contribution of this thesis is to bridge the gap between the zero memory PCFB and the infinite-order one. To that end, a variety of methods for the design of realizable signal-adapted FIR filter banks is presented. It is shown that a popular conventional method for designing signal-adapted FIR PU filter banks, which only requires the design of an optimal FIR compaction filter, is in fact not well suited for designing good filter banks due to the exponential complexity caused by the nonuniqueness of the FIR compaction filter. To avoid this dilemma, we propose a method by which all of the filters are obtained together. In particular, the method consists of finding an FIR PU least-squares approximant to the infinite-order PCFB polyphase matrix. Using an elegant complete parameterization of FIR PU systems in terms of canonical building blocks, an iterative greedy algorithm for solving the least-squares problem is presented. Simulation results provided here show that as the order or memory of the signal-adapted FIR PU filter bank increases, the filter bank behaves more and more like the infinite-order PCFB in terms of a variety of objectives including coding gain, multiresolution, and power minimization. This serves to bridge the gap between the zero memory and infinite memory PCFBs, which previously has not been done in the literature.
\r\n\r\nIn addition to being useful for the design of PCFB-like FIR filter banks, the proposed iterative algorithm can also be used for a variety of other design problems including the FIR PU interpolation problem. Unlike the traditional FIR interpolation problem, whose solution is known in closed form, the FIR PU interpolation problem is far more difficult and is in fact still open. Despite this, the proposed algorithm can be used to find an approximant to an interpolant and sometimes even find an interpolant, as we show here through simulations.
\r\n\r\nIn the second part of the thesis, we focus on the design of realizable signal-adapted quantized filter banks in which the filters are FIR but otherwise unconstrained. The filters are chosen to minimize the mean-squared error of the output, which is shown to be equivalent to maximizing the coding gain of the system. By alternately optimizing the analysis and synthesis filters, an iterative greedy algorithm, different from that mentioned above, is proposed for the design of such filter banks. Simulation results provided show that the filter banks designed exhibit performance close to the information theoretic rate-distortion bound.
\r\n\r\nFinally, we show how some of the techniques used in the above iterative algorithms can be used for the design of a channel shortening equalizer. Channel shortening equalizers, which arise in the context of digital communications, have been found to be necessary for DMT systems such as the digital subscriber loop (DSL) in which the channel impulse response must be shortened to the length of the cyclic prefix. In particular, we show how the eigenfilter technique which is used in the above-mentioned FIR PU iterative greedy algorithm, can be used for the design of a noise optimized channel shortening equalizer. As opposed to other techniques, which require a Cholesky decomposition of a certain matrix for every delay parameter considered, the proposed method is lower in complexity in that it only requires a single such decomposition for all delay values. Despite this significant decrease in complexity, it is shown through simulations that the equalizers designed using this technique perform nearly optimally in terms of observed bit rate.
", "doi": "10.7907/QZYN-ND12", "publication_date": "2004", "thesis_type": "phd", "thesis_year": "2004" }, { "id": "thesis:2381", "collection": "thesis", "collection_id": "2381", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-06022004-013457", "primary_object_url": { "basename": "Diego_Dugatkin_Thesis.pdf", "content": "final", "filesize": 1007427, "license": "other", "mime_type": "application/pdf", "url": "/2381/1/Diego_Dugatkin_Thesis.pdf", "version": "v2.0.0" }, "type": "thesis", "title": "Optimization of Multi-Resolution Source Codes", "author": [ { "family_name": "Dugatkin", "given_name": "Diego G.", "clpid": "Dugatkin-Diego-G" } ], "thesis_advisor": [ { "family_name": "Effros", "given_name": "Michelle", "clpid": "Effros-M" } ], "thesis_committee": [ { "family_name": "Effros", "given_name": "Michelle", "clpid": "Effros-M" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Vaidyanathan", "given_name": "P. P.", "clpid": "Vaidyanathan-P-P" }, { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" }, { "family_name": "Chan", "given_name": "Tony F-C", "clpid": "Chan-T-F" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "This thesis studies the optimization of multi-resolution source codes. A multi-resolution source code is a data compression algorithm that generates a bit-stream that can be truncated at any point to reconstruct low-resolution representations of the original data. By progressively refining the description, these codes allow the receiver to get representations of progressively increasing quality from a single file.
\r\n\r\nThe optimization methods presented here are based on the minimization of a Lagrangian performance measure, which is a weighted sum of rates and distortions at the different resolutions of the multi-resolution code. The Lagrangian coefficients are the weights that parameterize the priorities assigned to the resolutions. The relative value of these parameters can be set according to the user's preferences regarding which rates are more important, the probability of decoding the file at each possible rate, or any other prioritization rationale. We present a method for converting design constraints into the corresponding Lagrangian parameters.
\r\n\r\nWe also use a Lagrangian analysis to investigate optimality properties of multi-resolution codes. Specifically, we explore the characterization of the theoretically optimal output density functions of a two-resolution source code for any arbitrary set of priorities over the resolutions.
\r\n\r\nOnce the priority function has been identified, the goal is to design the multi-resolution code that yields the best rate-distortion trade-off for those priorities. The minimization of the multi-resolution Lagrangian is somewhat specific to the framework and type of multi-resolution code. We pursue this goal in several coding frameworks.
\r\n\r\nThe first framework is the multi-resolution vector quantizer (MRVQ) framework. Prior work on the topic described optimal MRVQ design for both fixed- and variable-rate systems but implemented only fixed-rate codes. The earliest portion of this thesis began with the implementation of the earlier described algorithm for variable-rate MRVQ for use as a testbed for understanding the important question of how to choose the Lagrangian parameters for multi-resolution codes to meet a collection of desired constraints.
\r\n\r\nArmed with a new understanding of parameter choice in the MRVQ framework, we moved next to the more sophisticated coding framework of wavelet-based embedded bit-plane coders. New results in this framework include improvements on the Set Partitioning in Hierarchical Trees (SPIHT) and the Group Testing for Wavelets (GTW) algorithms that apply the lessons learned from MRVQ theory in these more sophisticated wavelet coding frameworks. Experimental results demonstrate the performance benefits associated with this approach.
", "doi": "10.7907/RZ2Y-1Z78", "publication_date": "2004", "thesis_type": "phd", "thesis_year": "2004" }, { "id": "thesis:1909", "collection": "thesis", "collection_id": "1909", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-05212004-015508", "primary_object_url": { "basename": "Thesis_DaiLu.pdf", "content": "final", "filesize": 7365248, "license": "other", "mime_type": "application/pdf", "url": "/1909/1/Thesis_DaiLu.pdf", "version": "v2.0.0" }, "type": "thesis", "title": "Active Patch Array Design and Indoor Channel Modeling for Future Wireless Communications", "author": [ { "family_name": "Lu", "given_name": "Dai", "clpid": "Lu-Dai" } ], "thesis_advisor": [ { "family_name": "Rutledge", "given_name": "David B.", "clpid": "Rutledge-D-B" } ], "thesis_committee": [ { "family_name": "Rutledge", "given_name": "David B.", "clpid": "Rutledge-D-B" }, { "family_name": "Hajimiri", "given_name": "Ali", "clpid": "Hajimiri-A" }, { "family_name": "Effros", "given_name": "Michelle", "clpid": "Effros-M" }, { "family_name": "Pogorzelski", "given_name": "Ronald", "clpid": "Pogorzelski-R" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "Indoor wireless LAN systems currently operate at ranges of up to 30 meters, with practical data rates of 10 Mbps. In order to achieve higher data rates, higher frequencies are under consideration. Smaller antennas are required at these frequencies, but path loss increases. To combat the large path loss and multipath, 24 GHz phased arrays are being considered. The advantage of a phased array is that it can form narrow beams to favored directions, and nulls to combat interference. A 24 GHz active antenna with a 5-element patch array is demonstrated that includes an integrated GaAs MMIC power amplifier and low noise amplifier chip. Bias switching is used for changing from transmit to receive. The measured active gain is 31 dB in receive and 35 dB in transmit. The measured noise figure in receive is 3.5 dB and the maximum output power in transmit is 22 dBm (158 mW). Indoor wireless channels are investigated at five different frequency bands. The understanding of the channel will help link budgeting and system planning for future wireless communication. An automatic testing system has been developed using remote control by LABVIEW. This increases testing efficiency and reduces near field interference from the operator. A combined E/H plane 2-D ray-tracing method is proposed to predict the channel performance. This approach accurately predicts path loss for both line-of-sight and non-line-of-sight paths. It predicts the delay spread in line-of-sight paths well but fails for non-line-of-sight paths. This could be due to the ignorance of some higher order paths with small amplitudes but near random phases. In addition, a 3-D simplified ray-tracing code is developed to for access point optimization and to predict human shadow effects.\r\n", "doi": "10.7907/PVB9-TH19", "publication_date": "2004", "thesis_type": "phd", "thesis_year": "2004" }, { "id": "thesis:2455", "collection": "thesis", "collection_id": "2455", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-06052004-180113", "primary_object_url": { "basename": "FinalThesis.pdf", "content": "final", "filesize": 5372636, "license": "other", "mime_type": "application/pdf", "url": "/2455/1/FinalThesis.pdf", "version": "v3.0.0" }, "type": "thesis", "title": "Frequency Generation Techniques for Integrated Applications", "author": [ { "family_name": "Aparicio Joo", "given_name": "Roberto", "orcid": "0000-0003-2910-9969", "clpid": "Aparicio-Joo-Roberto" } ], "thesis_advisor": [ { "family_name": "Hajimiri", "given_name": "Ali", "clpid": "Hajimiri-A" } ], "thesis_committee": [ { "family_name": "Hajimiri", "given_name": "Ali", "clpid": "Hajimiri-A" }, { "family_name": "Tai", "given_name": "Yu-Chong", "clpid": "Tai-Yu-Chong" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Vahala", "given_name": "Kerry J.", "clpid": "Vahala-K-J" }, { "family_name": "Weinreb", "given_name": "Sander", "clpid": "Weinreb-S" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "This thesis presents novel oscillator topologies and passive structures that demonstrate improvements in performance compared to existing devices in CMOS. The contributions of this work include the development of original topologies and concepts together with practical implications in the area of integrated frequency generation.
\r\n\r\nA noise-shifting differential Colpitts oscillator topology is proposed. It is less sensitive to noise generated by the active devices than commonly used integrated oscillator topologies such as NMOS- or PMOS-only, and complementary cross-coupled. This is achieved through cyclostationary noise alignment while providing a fully differential output and large loop gain for reliable start up. An optimization strategy is derived for this oscillator that is used in the implementation of a CMOS prototype. The performance of this oscillator is compared to traditional topologies and previously published integrated oscillators achieving lower phase noise and some of the highest figures of merit, respectively.
\r\n\r\nA new circular-geometry oscillator topology is introduced. It allows the implementation of slab inductors for high-frequency and low-phase noise oscillator applications. Slab inductors present an attractive alternative for monolitic applications where low loss, low impedance, and high self-resonance integrated inductors are required. A general methodology to ensure the proper oscillation mode when several oscillator cores are coupled in a circular-geometry as well as to achieve a stable dc bias point is offered. Several circular-geometry CMOS integrated oscillator prototypes are presented as a proof of concept and their performances are compared to previously published high frequency oscillators achieving some of the best figures of merit.
\r\n\r\nTheoretical limits for the capacitance density of integrated capacitors with combined lateral and vertical field components are derived. These limits are used to investigate the efficiency of various capacitive structures such as lateral flux and quasi-fractal capacitors. This study leads to two new capacitor structures with high lateral-field efficiencies. These new capacitors demonstrate larger capacities, superior matching properties, tighter tolerances, and higher self-resonance frequencies than the standard horizontal parallel plate and previously reported lateral-field capacitors, while maintaining comparable quality factors. These superior qualities are verified by simulation and experimental results.
\r\n\r\nFinally, three phase-locked-loops (PLL) are presented. A 6.6GHz PLL for applications in a concurrent dual-band CMOS receiver is described. Careful frequency planning allows the generation of the three local oscillator signals required by the entire receiver using only one PLL, reducing power consumption and chip area considerably. The design issues of an ultra-low-power PLL prototype implemented in a sub-micron CMOS process are also discussed. The design of a low-power 3.2GHz PLL implementing a phase-compensation technique for fractional-N frequency synthesis is described. It uses an on-chip delay-locked-loop tuning scheme that attenuates the fractional spur independent of the output frequency and process variations.
", "doi": "10.7907/S2EP-3A93", "publication_date": "2004", "thesis_type": "phd", "thesis_year": "2004" }, { "id": "thesis:1887", "collection": "thesis", "collection_id": "1887", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-05202003-154451", "primary_object_url": { "basename": "thesis.pdf", "content": "final", "filesize": 2571860, "license": "other", "mime_type": "application/pdf", "url": "/1887/1/thesis.pdf", "version": "v3.0.0" }, "type": "thesis", "title": "Wireless Networks, from Collective Behavior to the Physics of Propagation", "author": [ { "family_name": "Franceschetti", "given_name": "Massimo", "orcid": "0000-0002-4057-8152", "clpid": "Franceschetti-Massimo" } ], "thesis_advisor": [ { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" } ], "thesis_committee": [ { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" }, { "family_name": "Hassibi", "given_name": "Babak", "orcid": "0000-0002-1375-5838", "clpid": "Hassibi-B" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Schulman", "given_name": "Leonard J.", "orcid": "0000-0001-9901-2797", "clpid": "Schulman-L-J" }, { "family_name": "Abu-Mostafa", "given_name": "Yaser S.", "clpid": "Abu-Mostafa-Y-S" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "This thesis addresses some of the key challenges in the emerging wireless scenario. It focuses on the problems of connectivity, coverage, and wave propagation, following a mathematically rigorous approach. The questions addressed are very basic and extremely easy to state. Their solution, however, can be difficult and leads to the development of a new kind of percolation theory, to a new theorem in geometry, and to a new model of wave propagation in urban environments. The problems are connected together to provide guidelines in the design of wireless networks.\r\n", "doi": "10.7907/SCTG-FN57", "publication_date": "2003", "thesis_type": "phd", "thesis_year": "2003" }, { "id": "thesis:2652", "collection": "thesis", "collection_id": "2652", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-06202002-170522", "primary_object_url": { "basename": "thesis.pdf", "content": "final", "filesize": 571645, "license": "other", "mime_type": "application/pdf", "url": "/2652/1/thesis.pdf", "version": "v3.0.0" }, "type": "thesis", "title": "Graph-Based Codes and Iterative Decoding", "author": [ { "family_name": "Khandekar", "given_name": "Aamod Dinkar", "clpid": "Khandekar-Aamod-Dinkar" } ], "thesis_advisor": [ { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" } ], "thesis_committee": [ { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" }, { "family_name": "Hassibi", "given_name": "Babak", "orcid": "0000-0002-1375-5838", "clpid": "Hassibi-B" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Preskill", "given_name": "John P.", "orcid": "0000-0002-2421-4762", "clpid": "Preskill-J" }, { "family_name": "Vaidyanathan", "given_name": "P. P.", "orcid": "0000-0003-3003-7042", "clpid": "Vaidyanathan-P-P" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "The field of error correcting codes was revolutionized by the introduction of turbo codes in 1993. These codes demonstrated dramatic performance improvements over any previously known codes, with significantly lower complexity. Since then, much progress has been made towards understanding the performance of these codes, as well as in using this understanding to design even better codes.
\r\n\r\nThis thesis takes a few more steps in both these directions. We develop a new technique, called the typical set bound, for analyzing the asymptotic performance of code ensembles based on their weight enumerators. This technique yields very tight bounds on the maximum-likelihood decoding threshold of code ensembles, and is powerful enough to reproduce Shannon's noisy coding theorem for the class of binary-input symmetric channels.
\r\n\r\nWe also introduce a new class of codes called irregular repeat-accumulate (IRA) codes, which are adapted from the previously known class of repeat-accumulate (RA) codes. These codes are competitive in terms of decoding performance with the class of irregular low-density parity-check (LDPC) codes, which are arguably the best class of codes known today, at least for long block lengths. In addition, IRA codes have a significant advantage over irregular LDPC codes in terms of encoding complexity.
\r\n\r\nWe also derive an analytical bound regarding iterative decoding thresholds of code ensembles on general binary-input symmetric channels, an area in which theoretical results are currently lacking.
", "doi": "10.7907/Q06G-MW38", "publication_date": "2003", "thesis_type": "phd", "thesis_year": "2003" }, { "id": "thesis:2652", "collection": "thesis", "collection_id": "2652", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-06202002-170522", "primary_object_url": { "basename": "thesis.pdf", "content": "final", "filesize": 571645, "license": "other", "mime_type": "application/pdf", "url": "/2652/1/thesis.pdf", "version": "v3.0.0" }, "type": "thesis", "title": "Graph-Based Codes and Iterative Decoding", "author": [ { "family_name": "Khandekar", "given_name": "Aamod Dinkar", "clpid": "Khandekar-Aamod-Dinkar" } ], "thesis_advisor": [ { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" } ], "thesis_committee": [ { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" }, { "family_name": "Hassibi", "given_name": "Babak", "orcid": "0000-0002-1375-5838", "clpid": "Hassibi-B" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Preskill", "given_name": "John P.", "orcid": "0000-0002-2421-4762", "clpid": "Preskill-J" }, { "family_name": "Vaidyanathan", "given_name": "P. P.", "orcid": "0000-0003-3003-7042", "clpid": "Vaidyanathan-P-P" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "The field of error correcting codes was revolutionized by the introduction of turbo codes in 1993. These codes demonstrated dramatic performance improvements over any previously known codes, with significantly lower complexity. Since then, much progress has been made towards understanding the performance of these codes, as well as in using this understanding to design even better codes.
\r\n\r\nThis thesis takes a few more steps in both these directions. We develop a new technique, called the typical set bound, for analyzing the asymptotic performance of code ensembles based on their weight enumerators. This technique yields very tight bounds on the maximum-likelihood decoding threshold of code ensembles, and is powerful enough to reproduce Shannon's noisy coding theorem for the class of binary-input symmetric channels.
\r\n\r\nWe also introduce a new class of codes called irregular repeat-accumulate (IRA) codes, which are adapted from the previously known class of repeat-accumulate (RA) codes. These codes are competitive in terms of decoding performance with the class of irregular low-density parity-check (LDPC) codes, which are arguably the best class of codes known today, at least for long block lengths. In addition, IRA codes have a significant advantage over irregular LDPC codes in terms of encoding complexity.
\r\n\r\nWe also derive an analytical bound regarding iterative decoding thresholds of code ensembles on general binary-input symmetric channels, an area in which theoretical results are currently lacking.
", "doi": "10.7907/Q06G-MW38", "publication_date": "2003", "thesis_type": "phd", "thesis_year": "2003" }, { "id": "thesis:1887", "collection": "thesis", "collection_id": "1887", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-05202003-154451", "primary_object_url": { "basename": "thesis.pdf", "content": "final", "filesize": 2571860, "license": "other", "mime_type": "application/pdf", "url": "/1887/1/thesis.pdf", "version": "v3.0.0" }, "type": "thesis", "title": "Wireless Networks, from Collective Behavior to the Physics of Propagation", "author": [ { "family_name": "Franceschetti", "given_name": "Massimo", "orcid": "0000-0002-4057-8152", "clpid": "Franceschetti-Massimo" } ], "thesis_advisor": [ { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" } ], "thesis_committee": [ { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" }, { "family_name": "Hassibi", "given_name": "Babak", "orcid": "0000-0002-1375-5838", "clpid": "Hassibi-B" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Schulman", "given_name": "Leonard J.", "orcid": "0000-0001-9901-2797", "clpid": "Schulman-L-J" }, { "family_name": "Abu-Mostafa", "given_name": "Yaser S.", "clpid": "Abu-Mostafa-Y-S" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "This thesis addresses some of the key challenges in the emerging wireless scenario. It focuses on the problems of connectivity, coverage, and wave propagation, following a mathematically rigorous approach. The questions addressed are very basic and extremely easy to state. Their solution, however, can be difficult and leads to the development of a new kind of percolation theory, to a new theorem in geometry, and to a new model of wave propagation in urban environments. The problems are connected together to provide guidelines in the design of wireless networks.\r\n", "doi": "10.7907/SCTG-FN57", "publication_date": "2003", "thesis_type": "phd", "thesis_year": "2003" }, { "id": "thesis:2289", "collection": "thesis", "collection_id": "2289", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-05302003-125004", "primary_object_url": { "basename": "Zhao_Qian_thesis.pdf", "content": "final", "filesize": 4131938, "license": "other", "mime_type": "application/pdf", "url": "/2289/1/Zhao_Qian_thesis.pdf", "version": "v3.0.0" }, "type": "thesis", "title": "Network Source Coding: Theory and Code Design for Broadcast and Multiple Access Networks", "author": [ { "family_name": "Zhao", "given_name": "Qian", "clpid": "Zhao-Qian" } ], "thesis_advisor": [ { "family_name": "Effros", "given_name": "Michelle", "orcid": "0000-0003-3757-0675", "clpid": "Effros-M" } ], "thesis_committee": [ { "family_name": "Effros", "given_name": "Michelle", "orcid": "0000-0003-3757-0675", "clpid": "Effros-M" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" }, { "family_name": "Low", "given_name": "Steven H.", "orcid": "0000-0001-6476-3048", "clpid": "Low-S-H" }, { "family_name": "Abu-Mostafa", "given_name": "Yaser S.", "clpid": "Abu-Mostafa-Y-S" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "In the information age, network systems and applications have been growing rapidly to provide us with more versatile and high bit rate services. However, the limited bandwidth restricts the amount of information that can be sent through the networks. Thus efficient data representation or source coding is imperative for future network development. Distinct from the traditional source coding strategy, network source codes take advantage of the network topology and are able to maximally compress data before transmission.
\r\n\r\nIn this thesis, I present a variety of source coding techniques for use in network environments and demonstrate the benefits of network source codes over traditional source codes from both theoretical and practical perspectives.
\r\n\r\nFirst, I address source coding for broadcast systems. The results I obtain include derivation of the theoretical limits of broadcast system source codes, algorithm design for optimal broadcast system vector quantizers, implementation of the optimal code, and experimental results.
\r\n\r\nThen, I focus on multiple access systems which are the dual systems of broadcast systems. I present the properties of multiple access source codes and generalize traditional entropy code design algorithms to attain the corresponding optimal multiple access source codes for arbitrary joint source statistics. I further introduce a family of polynomial complexity code design algorithms that approximates the optimal solutions. Application to universal coding for multiple access networks when the joint source statistics are unknown a priori is briefly discussed. Finally, I demonstrate algorithmic performance by showing experimental results on a variety of data sets.
\r\n\r\ninally, in seeking a simple lossy source coding method for general networks, I apply entropy constrained dithered quantization in network source code design and present the coding results for multi-resolution source codes and multiple access source codes. Multi-resolution and multiple access dithered quantizers are low complexity codes that achieve performance very close to the theoretical rate-distortion bound.
", "doi": "10.7907/61XN-MV62", "publication_date": "2003", "thesis_type": "phd", "thesis_year": "2003" }, { "id": "thesis:4917", "collection": "thesis", "collection_id": "4917", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-12102002-113833", "primary_object_url": { "basename": "thesis.pdf", "content": "final", "filesize": 1898611, "license": "other", "mime_type": "application/pdf", "url": "/4917/1/thesis.pdf", "version": "v4.0.0" }, "type": "thesis", "title": "A Probabilistic Approach to Human Motion Detection and Labeling", "author": [ { "family_name": "Song", "given_name": "Yang", "clpid": "Song-Yang" } ], "thesis_advisor": [ { "family_name": "Perona", "given_name": "Pietro", "clpid": "Perona-P" } ], "thesis_committee": [ { "family_name": "Perona", "given_name": "Pietro", "clpid": "Perona-P" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Murray", "given_name": "Richard M.", "clpid": "Murray-R-M" }, { "family_name": "Soatto", "given_name": "Stefano", "clpid": "Soatto-Stefano" }, { "family_name": "Abu-Mostafa", "given_name": "Yaser S.", "clpid": "Abu-Mostafa-Y-S" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "Human motion analysis is a very important task for computer vision with many potential applications. There are several problems in human motion analysis: detection, tracking, and activity interpretation. Detection is the most fundamental problem of the three, but remains untackled due to its inherent difficulty. This thesis develops a solution to the problem. It is based on a learned probabilistic model of the joint positions and velocities of the body parts, where detection and labeling are performed by hypothesis testing on the maximum a posterior estimate of the pose and motion of the body. To achieve efficiency in learning and testing, a graphical model is used to approximate the conditional independence of human motion. This model is also shown to provide a natural way to deal with clutter and occlusion.
\r\n\r\nOne key factor in the proposed method is the probabilistic model of human motion. In this thesis, an unsupervised learning algorithm that can obtain the probabilistic model automatically from unlabeled training data is presented. The training data include useful foreground features as well as features that arise from irrelevant background clutter. The correspondence between parts and detected features is also unknown in the training data. To learn the best model structure as well as model parameters, a variant of the EM algorithm is developed where the labeling of the data (part assignments) is treated as hidden variables. We explore two classes of graphical models: trees and decomposable triangulated graphs and find that the later are superior for our application. To better model human motion, we also consider the case when the model consists of mixtures of decomposable triangulated graphs.
\r\n \r\nThe efficiency and effectiveness of the algorithm have been demonstrated by applying it to generate models of human motion automatically from unlabeled image sequences, and testing the learned models on a variety of sequences. We find detection rates of over 95% on pairs of frames. This is very promising for building a real-life system, for example, a pedestrian detector.
", "doi": "10.7907/945J-QX86", "publication_date": "2003", "thesis_type": "phd", "thesis_year": "2003" }, { "id": "thesis:748", "collection": "thesis", "collection_id": "748", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-02252003-134943", "primary_object_url": { "basename": "david_randall_kent_iv-dissertation.pdf", "content": "final", "filesize": 627928, "license": "other", "mime_type": "application/pdf", "url": "/748/1/david_randall_kent_iv-dissertation.pdf", "version": "v2.0.0" }, "type": "thesis", "title": "New Quantum Monte Carlo Algorithms to Efficiently Utilize Massively Parallel Computers", "author": [ { "family_name": "Kent", "given_name": "David Randall, IV", "clpid": "Kent-David-Randall-IV" } ], "thesis_advisor": [ { "family_name": "Gray", "given_name": "Harry B.", "clpid": "Gray-H-B" }, { "family_name": "Goddard", "given_name": "William A., III", "clpid": "Goddard-W-A-III" } ], "thesis_committee": [ { "family_name": "Blake", "given_name": "Geoffrey A.", "clpid": "Blake-G-A" }, { "family_name": "Gray", "given_name": "Harry B.", "clpid": "Gray-H-B" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Lewis", "given_name": "Nathan Saul", "clpid": "Lewis-N-S" }, { "family_name": "Goddard", "given_name": "William A., III", "clpid": "Goddard-W-A-III" } ], "local_group": [ { "literal": "div_chem" } ], "abstract": "The exponential growth in computer power over the past few decades has been a huge boon to computational chemistry, physics, biology, and materials science. Now, a standard workstation or Linux cluster can calculate semi-quantitative properties of moderately sized systems. The next step in computational science is developing better algorithms which allow quantitative calculations of a system's properties.
\r\n\r\nA relatively new class of algorithms, known collectively as Quantum Monte Carlo (QMC), has the potential to quantitatively calculate the properties of molecular systems. Furthermore, QMC scales as O(N\u00b3) or better. This makes possible very high-level calculations on systems that are too large to be examined using standard high-level methods.
\r\n\r\nThis thesis develops (1) an efficient algorithm for determining \"on-the-fly\" the statistical error in serially correlated data, (2) a manager-worker parallelization algorithm for QMC that allows calculations to run on heterogeneous parallel computers and computational grids, (3) a robust algorithm for optimizing Jastrow functions which have singularities for some parameter values, and (4) a proof-of-concept demonstrating that it is possible to find transferable parameter sets for large classes of compounds.
", "doi": "10.7907/V64A-V618", "publication_date": "2003", "thesis_type": "phd", "thesis_year": "2003" }, { "id": "thesis:1777", "collection": "thesis", "collection_id": "1777", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-05132005-151145", "primary_object_url": { "basename": "Foltz_k_2002.pdf", "content": "final", "filesize": 703084, "license": "other", "mime_type": "application/pdf", "url": "/1777/1/Foltz_k_2002.pdf", "version": "v2.0.0" }, "type": "thesis", "title": "Periodic Broadcast Scheduling for Data Distribution", "author": [ { "family_name": "Foltz", "given_name": "Kevin E.", "clpid": "Foltz-Kevin-E" } ], "thesis_advisor": [ { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" } ], "thesis_committee": [ { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Effros", "given_name": "Michelle", "orcid": "0000-0003-3757-0675", "clpid": "Effros-M" }, { "family_name": "Hassibi", "given_name": "Babak", "orcid": "0000-0002-1375-5838", "clpid": "Hassibi-B" }, { "family_name": "Low", "given_name": "Steven H.", "orcid": "0000-0001-6476-3048", "clpid": "Low-S-H" }, { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "As wireless computer networks grow in size and complexity, we are faced with the problem of providing scalable, high-bandwidth service to their users. Wired networks typically use \"data pull,\" where users send requests to a server and the server responds with the desired information. In the wireless domain, \"data push\" promises to provide better performance for many applications [1]. The broadcast domain that is typical of wireless communication is very effective in distributing information to large audiences.\r\n\r\nThe idea of broadcast disks has been around since the Teletext system [3]. There is now an interest in applying these ideas to wireless computer networks. There are some interesting research questions about scheduling for data distribution. Computing optimal schedules has been shown to be difficult [18]. The optimal schedules themselves, however, seem to be less complex, and often periodic [4]. Xu [24] looks at the scheduling of streaming data, which involves splitting the data into smaller pieces. The idea of error correction is also important for wireless transmission due to the noisy nature of the channel [6].\r\n\r\nWe look at scheduling data for broadcast. We compare time-division scheduling and frequency-division scheduling for data items of equal length. We show that time-division is better for sending dynamic data. We then find optimal time-division schedules for two items. We show how the freedom to split items into smaller pieces can give improvements in performance. With a single split, where each of two items is split in half, we find the optimal schedules for items of equal length.\r\n\r\nWe continue with the idea of splitting items, and show what happens when the number of splits is very large. Then, we examine what happens when we add streaming data to our broadcast. We compare time-division and frequency-division as before, and now also look at a mix of the two. We prove bounds on where the mix is the best broadcast method.", "doi": "10.7907/980Q-RQ20", "publication_date": "2002", "thesis_type": "phd", "thesis_year": "2002" }, { "id": "thesis:3347", "collection": "thesis", "collection_id": "3347", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-09062005-083717", "primary_object_url": { "basename": "Nicholson_a_2002.pdf", "content": "final", "filesize": 3500805, "license": "other", "mime_type": "application/pdf", "url": "/3347/1/Nicholson_a_2002.pdf", "version": "v2.0.0" }, "type": "thesis", "title": "Generalization Error Estimates and Training Data Valuation", "author": [ { "family_name": "Nicholson", "given_name": "Alexander Marshall", "clpid": "Nicholson-Alexander-Marshall" } ], "thesis_advisor": [ { "family_name": "Abu-Mostafa", "given_name": "Yaser S.", "clpid": "Abu-Mostafa-Y-S" } ], "thesis_committee": [ { "family_name": "Abu-Mostafa", "given_name": "Yaser S.", "clpid": "Abu-Mostafa-Y-S" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Hickey", "given_name": "Jason J.", "clpid": "Hickey-J-J" }, { "family_name": "Magdon-Ismail", "given_name": "Malik", "clpid": "Magdon-Ismail" }, { "family_name": "Perona", "given_name": "Pietro", "orcid": "0000-0002-7583-5809", "clpid": "Perona-P" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "This thesis addresses several problems related to generalization in machine learning systems. We introduce a theoretical framework for studying learning and generalization. Within this framework, a closed form is derived for the expected generalization error that estimates the out-of-sample performance in terms of the in-sample performance. We consider the problem of overfitting and show that, using a simple exhaustive learning algorithm, overfitting does not occur. These results do not assume a particular form of the target function, input distribution or learning model, and hold even with noisy data sets. We apply our analysis to practical learning systems, illustrate how it may be used to estimate out-of-sample errors in practice, and demonstrate that the resulting estimates improve upon errors estimated with a validation set for real world problems.\r\n\r\nBased on this study of generalization, we develop a technique for quantitative valuation of training data. We demonstrate that this valuation may be used to select training sets that improve generalization performance. With a reasonable prior over target functions, it further allows us to estimate the level of noise in a data set and provides for detection and correction of noise in individual examples. Finally, this data valuation can be used to classify new examples, yielding a new learning algorithm that is shown to be relatively robust to noise.\r\n", "doi": "10.7907/1H16-VX81", "publication_date": "2002", "thesis_type": "phd", "thesis_year": "2002" }, { "id": "thesis:3347", "collection": "thesis", "collection_id": "3347", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-09062005-083717", "primary_object_url": { "basename": "Nicholson_a_2002.pdf", "content": "final", "filesize": 3500805, "license": "other", "mime_type": "application/pdf", "url": "/3347/1/Nicholson_a_2002.pdf", "version": "v2.0.0" }, "type": "thesis", "title": "Generalization Error Estimates and Training Data Valuation", "author": [ { "family_name": "Nicholson", "given_name": "Alexander Marshall", "clpid": "Nicholson-Alexander-Marshall" } ], "thesis_advisor": [ { "family_name": "Abu-Mostafa", "given_name": "Yaser S.", "clpid": "Abu-Mostafa-Y-S" } ], "thesis_committee": [ { "family_name": "Abu-Mostafa", "given_name": "Yaser S.", "clpid": "Abu-Mostafa-Y-S" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Hickey", "given_name": "Jason J.", "clpid": "Hickey-J-J" }, { "family_name": "Magdon-Ismail", "given_name": "Malik", "clpid": "Magdon-Ismail" }, { "family_name": "Perona", "given_name": "Pietro", "orcid": "0000-0002-7583-5809", "clpid": "Perona-P" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "This thesis addresses several problems related to generalization in machine learning systems. We introduce a theoretical framework for studying learning and generalization. Within this framework, a closed form is derived for the expected generalization error that estimates the out-of-sample performance in terms of the in-sample performance. We consider the problem of overfitting and show that, using a simple exhaustive learning algorithm, overfitting does not occur. These results do not assume a particular form of the target function, input distribution or learning model, and hold even with noisy data sets. We apply our analysis to practical learning systems, illustrate how it may be used to estimate out-of-sample errors in practice, and demonstrate that the resulting estimates improve upon errors estimated with a validation set for real world problems.\r\n\r\nBased on this study of generalization, we develop a technique for quantitative valuation of training data. We demonstrate that this valuation may be used to select training sets that improve generalization performance. With a reasonable prior over target functions, it further allows us to estimate the level of noise in a data set and provides for detection and correction of noise in individual examples. Finally, this data valuation can be used to classify new examples, yielding a new learning algorithm that is shown to be relatively robust to noise.\r\n", "doi": "10.7907/1H16-VX81", "publication_date": "2002", "thesis_type": "phd", "thesis_year": "2002" }, { "id": "thesis:4821", "collection": "thesis", "collection_id": "4821", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-12072001-160019", "primary_object_url": { "basename": "thesis-online.pdf", "content": "final", "filesize": 850716, "license": "other", "mime_type": "application/pdf", "url": "/4821/1/thesis-online.pdf", "version": "v3.0.0" }, "type": "thesis", "title": "Dynamic UNITY", "author": [ { "family_name": "Zimmerman", "given_name": "Daniel Marc", "clpid": "Zimmerman-Daniel-Marc" } ], "thesis_advisor": [ { "family_name": "Chandy", "given_name": "K. Mani", "clpid": "Chandy-K-M" } ], "thesis_committee": [ { "family_name": "Chandy", "given_name": "K. Mani", "clpid": "Chandy-K-M" }, { "family_name": "Martin", "given_name": "Alain J.", "clpid": "Martin-A-J" }, { "family_name": "Hickey", "given_name": "Jason J.", "clpid": "Hickey-J-J" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "Dynamic distributed systems, where a changing set of communicating processes must interoperate to accomplish particular computational tasks, are becoming extremely important. Designing and implementing these systems, and verifying the correctness of the designs and implementations, are difficult tasks. The goal of this thesis is to make these tasks easier.\r\n\r\nThis thesis presents a specification language for dynamic distributed systems, based on Chandy and Misra's UNITY language. It extends the UNITY language to enable process creation, process deletion, and dynamic communication patterns.\r\n\r\nThe thesis defines an execution model for systems specified in this language, which leads to a proof logic similar to that of UNITY. While extending UNITY logic to correctly handle systems with dynamic behavior, this logic retains the familiar UNITY operators and most of the proof rules associated with them. \r\n\r\nThe thesis presents specifications for three example dynamic distributed systems to demonstrate the use of the specification language, and full correctness proofs for two of these systems and a partial correctness proof for the third to demonstrate the use of the proof logic. \r\n\r\nThe thesis details a method for determining whether a system in the specification language can be transformed into an implementation in a standard programming language, as well as a method for performing this transformation on those specifications that can. This guarantees a correct implementation for any specification that can be so transformed. \r\n", "doi": "10.7907/AC6E-WE21", "publication_date": "2002", "thesis_type": "phd", "thesis_year": "2002" }, { "id": "thesis:4821", "collection": "thesis", "collection_id": "4821", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-12072001-160019", "primary_object_url": { "basename": "thesis-online.pdf", "content": "final", "filesize": 850716, "license": "other", "mime_type": "application/pdf", "url": "/4821/1/thesis-online.pdf", "version": "v3.0.0" }, "type": "thesis", "title": "Dynamic UNITY", "author": [ { "family_name": "Zimmerman", "given_name": "Daniel Marc", "clpid": "Zimmerman-Daniel-Marc" } ], "thesis_advisor": [ { "family_name": "Chandy", "given_name": "K. Mani", "clpid": "Chandy-K-M" } ], "thesis_committee": [ { "family_name": "Chandy", "given_name": "K. Mani", "clpid": "Chandy-K-M" }, { "family_name": "Martin", "given_name": "Alain J.", "clpid": "Martin-A-J" }, { "family_name": "Hickey", "given_name": "Jason J.", "clpid": "Hickey-J-J" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "Dynamic distributed systems, where a changing set of communicating processes must interoperate to accomplish particular computational tasks, are becoming extremely important. Designing and implementing these systems, and verifying the correctness of the designs and implementations, are difficult tasks. The goal of this thesis is to make these tasks easier.\r\n\r\nThis thesis presents a specification language for dynamic distributed systems, based on Chandy and Misra's UNITY language. It extends the UNITY language to enable process creation, process deletion, and dynamic communication patterns.\r\n\r\nThe thesis defines an execution model for systems specified in this language, which leads to a proof logic similar to that of UNITY. While extending UNITY logic to correctly handle systems with dynamic behavior, this logic retains the familiar UNITY operators and most of the proof rules associated with them. \r\n\r\nThe thesis presents specifications for three example dynamic distributed systems to demonstrate the use of the specification language, and full correctness proofs for two of these systems and a partial correctness proof for the third to demonstrate the use of the proof logic. \r\n\r\nThe thesis details a method for determining whether a system in the specification language can be transformed into an implementation in a standard programming language, as well as a method for performing this transformation on those specifications that can. This guarantees a correct implementation for any specification that can be so transformed. \r\n", "doi": "10.7907/AC6E-WE21", "publication_date": "2002", "thesis_type": "phd", "thesis_year": "2002" }, { "id": "thesis:2129", "collection": "thesis", "collection_id": "2129", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-05272004-094928", "primary_object_url": { "basename": "Aoki_i_2002.pdf", "content": "final", "filesize": 26641555, "license": "other", "mime_type": "application/pdf", "url": "/2129/1/Aoki_i_2002.pdf", "version": "v3.0.0" }, "type": "thesis", "title": "Distributed Active Transformer for Integrated Power Amplification", "author": [ { "family_name": "Aoki", "given_name": "Ichiro", "clpid": "Aoki-Ichiro" } ], "thesis_advisor": [ { "family_name": "Janda", "given_name": "Kenneth C.", "clpid": "Janda-K-C" } ], "thesis_committee": [ { "family_name": "Rutledge", "given_name": "David B.", "clpid": "Rutledge-D-B" }, { "family_name": "Hajimiri", "given_name": "Ali", "orcid": "0000-0001-6736-8019", "clpid": "Hajimiri-A" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Weinreb", "given_name": "Sander", "orcid": "0000-0002-9353-6204", "clpid": "Weinreb-S" }, { "family_name": "Tai", "given_name": "Yu-Chong", "orcid": "0000-0001-8529-106X", "clpid": "Tai-Yu-Chong" }, { "family_name": "Janda", "given_name": "Kenneth C.", "clpid": "Janda-K-C" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "A novel on-chip impedance transformation and power-combining technique, the distributed active transformer (DAT) is introduced. It overcomes the fundamental difficulties presented by silicon technology in the design of integrated rf power amplifiers. This technique efficiently combines several low-voltage push-pull amplifiers and simultaneously performs an impedance transformation to produce a larger output power while maintaining a 50\u03a9 match. It also uses virtual ac grounds and magnetic couplings extensively to eliminate the need for any off-chip component, such as tuned bonding wires or external inductors. Furthermore, it desensitizes the operation of the amplifier to the inductance of bonding wires making the design more repeatable.
\r\n\r\nIn this work, the performance of the introduced DAT structure is compared to that of conventional on-chip impedance transformation methods. Their fundamental power-efficiency limitations in the design of high-power fully-integrated amplifiers in standard silicon process technologies are analyzed and the DAT is demonstrated to be more efficient. Furthermore, different classes of power amplification operations and their use in DAT power amplifiers are studied.
\r\n\r\nTo demonstrate the feasibility of this concept several silicon integrated power amplifiers have been fabricated and measured including a 2.4-GHz, 2-W, 2-V truly fully-integrated power amplifier with 50\u03a9 on-chip input and output matching using 0.35\u00b5m CMOS transistors. It achieves a power added efficiency (PAE) of 41 % at this power level, demonstrating for the first time a truly fully-integrated watt-level GHz range CMOS power amplifier. It can also produce 450mW using a 1 V supply. A two stage DAT prototype, also at 2.4GHz using the same technology, operates with higher gain and lower supply voltage achieving 1-W output power, 30% PAE, and 14-dB gain with 1.15-V supply.
", "doi": "10.7907/6XMD-MR86", "publication_date": "2002", "thesis_type": "phd", "thesis_year": "2002" }, { "id": "thesis:2129", "collection": "thesis", "collection_id": "2129", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-05272004-094928", "primary_object_url": { "basename": "Aoki_i_2002.pdf", "content": "final", "filesize": 26641555, "license": "other", "mime_type": "application/pdf", "url": "/2129/1/Aoki_i_2002.pdf", "version": "v3.0.0" }, "type": "thesis", "title": "Distributed Active Transformer for Integrated Power Amplification", "author": [ { "family_name": "Aoki", "given_name": "Ichiro", "clpid": "Aoki-Ichiro" } ], "thesis_advisor": [ { "family_name": "Janda", "given_name": "Kenneth C.", "clpid": "Janda-K-C" } ], "thesis_committee": [ { "family_name": "Rutledge", "given_name": "David B.", "clpid": "Rutledge-D-B" }, { "family_name": "Hajimiri", "given_name": "Ali", "orcid": "0000-0001-6736-8019", "clpid": "Hajimiri-A" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Weinreb", "given_name": "Sander", "orcid": "0000-0002-9353-6204", "clpid": "Weinreb-S" }, { "family_name": "Tai", "given_name": "Yu-Chong", "orcid": "0000-0001-8529-106X", "clpid": "Tai-Yu-Chong" }, { "family_name": "Janda", "given_name": "Kenneth C.", "clpid": "Janda-K-C" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "A novel on-chip impedance transformation and power-combining technique, the distributed active transformer (DAT) is introduced. It overcomes the fundamental difficulties presented by silicon technology in the design of integrated rf power amplifiers. This technique efficiently combines several low-voltage push-pull amplifiers and simultaneously performs an impedance transformation to produce a larger output power while maintaining a 50\u03a9 match. It also uses virtual ac grounds and magnetic couplings extensively to eliminate the need for any off-chip component, such as tuned bonding wires or external inductors. Furthermore, it desensitizes the operation of the amplifier to the inductance of bonding wires making the design more repeatable.
\r\n\r\nIn this work, the performance of the introduced DAT structure is compared to that of conventional on-chip impedance transformation methods. Their fundamental power-efficiency limitations in the design of high-power fully-integrated amplifiers in standard silicon process technologies are analyzed and the DAT is demonstrated to be more efficient. Furthermore, different classes of power amplification operations and their use in DAT power amplifiers are studied.
\r\n\r\nTo demonstrate the feasibility of this concept several silicon integrated power amplifiers have been fabricated and measured including a 2.4-GHz, 2-W, 2-V truly fully-integrated power amplifier with 50\u03a9 on-chip input and output matching using 0.35\u00b5m CMOS transistors. It achieves a power added efficiency (PAE) of 41 % at this power level, demonstrating for the first time a truly fully-integrated watt-level GHz range CMOS power amplifier. It can also produce 450mW using a 1 V supply. A two stage DAT prototype, also at 2.4GHz using the same technology, operates with higher gain and lower supply voltage achieving 1-W output power, 30% PAE, and 14-dB gain with 1.15-V supply.
", "doi": "10.7907/6XMD-MR86", "publication_date": "2002", "thesis_type": "phd", "thesis_year": "2002" }, { "id": "thesis:6777", "collection": "thesis", "collection_id": "6777", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:01252012-135136531", "primary_object_url": { "basename": "Feldmann_mt_2002.pdf", "content": "final", "filesize": 53221379, "license": "other", "mime_type": "application/pdf", "url": "/6777/1/Feldmann_mt_2002.pdf", "version": "v5.0.0" }, "type": "thesis", "title": "Quantum Monte Carlo: Quest to Get Bigger, Faster, and Cheaper", "author": [ { "family_name": "Feldmann", "given_name": "Michael Todd", "clpid": "Feldmann-Michael-Todd" } ], "thesis_advisor": [ { "family_name": "Goddard", "given_name": "William A., III", "orcid": "0000-0003-0097-5716", "clpid": "Goddard-W-A-III" } ], "thesis_committee": [ { "family_name": "Kuppermann", "given_name": "Aron", "clpid": "Kuppermann-A" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Goddard", "given_name": "William A., III", "orcid": "0000-0003-0097-5716", "clpid": "Goddard-W-A-III" }, { "family_name": "Pierce", "given_name": "Niles A.", "orcid": "0000-0003-2367-4406", "clpid": "Pierce-N-A" }, { "family_name": "Grubbs", "given_name": "Robert H.", "orcid": "0000-0002-0057-7817", "clpid": "Grubbs-R-H" } ], "local_group": [ { "literal": "div_chem" } ], "abstract": "We reexamine some fundamental Quantum Monte Carlo (QMC) algorithms with the goal of making QMC more mainstream and efficient. Two major themes exist: (1) Make QMC faster and cheaper, and (2) Make QMC more robust and easier to use. A fast \"on-the-fly\" algorithm to extract uncorrelated estimators from serially correlated data on a huge network is presented, DDDA. A very efficient manager-worker algorithm for QMC parallelization is presented, QMC-MW. Reduced expense VMC optimization procedure is presented to better guess initial Jast row parameter sets for hydrocarbons, GJ. I also examine the formation and decomposition of aminomethanol using a variety of methods including a test of the hydrocarbon GJ set on these oxygen- and nitrogen-containing systems. The QMC program suite QMcBeaver is available from the authors in its entirety while a user's and developer's manual is attached as supplementary material.", "doi": "10.7907/4D4F-WZ34", "publication_date": "2002", "thesis_type": "phd", "thesis_year": "2002" }, { "id": "thesis:6777", "collection": "thesis", "collection_id": "6777", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:01252012-135136531", "primary_object_url": { "basename": "Feldmann_mt_2002.pdf", "content": "final", "filesize": 53221379, "license": "other", "mime_type": "application/pdf", "url": "/6777/1/Feldmann_mt_2002.pdf", "version": "v5.0.0" }, "type": "thesis", "title": "Quantum Monte Carlo: Quest to Get Bigger, Faster, and Cheaper", "author": [ { "family_name": "Feldmann", "given_name": "Michael Todd", "clpid": "Feldmann-Michael-Todd" } ], "thesis_advisor": [ { "family_name": "Goddard", "given_name": "William A., III", "orcid": "0000-0003-0097-5716", "clpid": "Goddard-W-A-III" } ], "thesis_committee": [ { "family_name": "Kuppermann", "given_name": "Aron", "clpid": "Kuppermann-A" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Goddard", "given_name": "William A., III", "orcid": "0000-0003-0097-5716", "clpid": "Goddard-W-A-III" }, { "family_name": "Pierce", "given_name": "Niles A.", "orcid": "0000-0003-2367-4406", "clpid": "Pierce-N-A" }, { "family_name": "Grubbs", "given_name": "Robert H.", "orcid": "0000-0002-0057-7817", "clpid": "Grubbs-R-H" } ], "local_group": [ { "literal": "div_chem" } ], "abstract": "We reexamine some fundamental Quantum Monte Carlo (QMC) algorithms with the goal of making QMC more mainstream and efficient. Two major themes exist: (1) Make QMC faster and cheaper, and (2) Make QMC more robust and easier to use. A fast \"on-the-fly\" algorithm to extract uncorrelated estimators from serially correlated data on a huge network is presented, DDDA. A very efficient manager-worker algorithm for QMC parallelization is presented, QMC-MW. Reduced expense VMC optimization procedure is presented to better guess initial Jast row parameter sets for hydrocarbons, GJ. I also examine the formation and decomposition of aminomethanol using a variety of methods including a test of the hydrocarbon GJ set on these oxygen- and nitrogen-containing systems. The QMC program suite QMcBeaver is available from the authors in its entirety while a user's and developer's manual is attached as supplementary material.", "doi": "10.7907/4D4F-WZ34", "publication_date": "2002", "thesis_type": "phd", "thesis_year": "2002" }, { "id": "thesis:1965", "collection": "thesis", "collection_id": "1965", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-05232002-173821", "primary_object_url": { "basename": "thesis.pdf", "content": "final", "filesize": 1064666, "license": "other", "mime_type": "application/pdf", "url": "/1965/1/thesis.pdf", "version": "v3.0.0" }, "type": "thesis", "title": "Rate Loss of Network Source Codes", "author": [ { "family_name": "Feng", "given_name": "Hanying", "clpid": "Feng-Hanying" } ], "thesis_advisor": [ { "family_name": "Effros", "given_name": "Michelle", "orcid": "0000-0003-3757-0675", "clpid": "Effros-M" } ], "thesis_committee": [ { "family_name": "Effros", "given_name": "Michelle", "orcid": "0000-0003-3757-0675", "clpid": "Effros-M" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" }, { "family_name": "Low", "given_name": "Steven H.", "orcid": "0000-0001-6476-3048", "clpid": "Low-S-H" }, { "family_name": "Abu-Mostafa", "given_name": "Yaser S.", "clpid": "Abu-Mostafa-Y-S" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "In this thesis, I present bounds on the performance of a variety of network source codes. These rate loss bounds compare the rates achievable by each network code to the rate-distortion bound R(D) at the corresponding distortions. The result is a collection of optimal performance bounds that are easy to calculate. \r\n\r\nI first present new bounds for the rate loss of multi-resolution source codes (MRSCs). Considering an M-resolution code with M>=2, the rate loss at the ith resolution with distortion D_i is defined as L_i=R_i-R(D_i), where R_i is the rate achievable by the MRSC at stage i. For 2-resolution codes, there are three scenarios of particular interest: (i) when both resolutions are equally important; (ii) when the rate loss at the first resolution is 0 (L_1=0); (iii) when the rate loss at the second resolution is 0 (L_2=0). The work of Lastras and Berger gives constant upper bounds for the rate loss in scenarios (i) and (ii) and an asymptotic bound for scenario (iii). In this thesis, I show a constant bound for scenario (iii), tighten the bounds for scenario (i) and (ii), and generalize the bound for scenario (ii) to M-resolution greedy codes. \r\n\r\nI also present upper bounds for the rate losses of additive MRSCs (AMRSCs), a special MRSC. I obtain two bounds on the rate loss of AMRSCs: one primarily good for low rate coding and another which depends on the source entropy. \r\n\r\nI then generalize the rate loss definition and present upper bounds for the rate losses of multiple description source codes. I divide the distortion region into three sub-regions and bound the rate losses by small constants in two sub-regions and by the joint rate losses of a normal source with the same variance in the other sub-region. \r\n\r\nFinally, I present bounds for the rate loss of multiple access source codes (MASCs). I show that lossy MASCs can be almost as good as codes based on joint source encoding.", "doi": "10.7907/GVDP-7248", "publication_date": "2002", "thesis_type": "phd", "thesis_year": "2002" }, { "id": "thesis:967", "collection": "thesis", "collection_id": "967", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-03162005-141416", "primary_object_url": { "basename": "Ham_d_2002.pdf", "content": "final", "filesize": 8325658, "license": "other", "mime_type": "application/pdf", "url": "/967/1/Ham_d_2002.pdf", "version": "v6.0.0" }, "type": "thesis", "title": "Statistical Electronics: Noise Processes in Integrated Communication Systems", "author": [ { "family_name": "Ham", "given_name": "Donhee", "clpid": "Ham-Donhee" } ], "thesis_advisor": [ { "family_name": "Hajimiri", "given_name": "Ali", "orcid": "0000-0001-6736-8019", "clpid": "Hajimiri-A" } ], "thesis_committee": [ { "family_name": "Hajimiri", "given_name": "Ali", "orcid": "0000-0001-6736-8019", "clpid": "Hajimiri-A" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Cross", "given_name": "Michael Clifford", "clpid": "Cross-M-C" }, { "family_name": "Rutledge", "given_name": "David B.", "clpid": "Rutledge-D-B" }, { "family_name": "Tai", "given_name": "Yu-Chong", "orcid": "0000-0001-8529-106X", "clpid": "Tai-Yu-Chong" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "This thesis presents a comprehensive investigation of noise and thermodynamics in electronic circuits and systems. This study of \"statistical electronics\" spans two disciplines, statistical thermodynamics and electronic circuit engineering, and leads to a general picture that bridges electronics and statistical thermodynamics.
\r\n\r\nOur work on statistical electronics has both scientific and engineering implications. Scientifically, this work is an extensive study of statistical thermodynamics in the context of electrical circuits, which has made several significant contributions to the understanding of noise processes in electrical circuits. The technological importance is a demonstration of how the fundamental physical considerations evolve to practical high-performance novel circuit design. The power of our fundamental approach is demonstrated through several practical circuit examples.
\r\n\r\nFirst, our investigation of fluctuations in nonlinear electrical circuits provides deep insight into the nonlinear fluctuation phenomena. Especially, the study of fluctuations in nonlinear active devices constitutes an important sector in this investigation; verifying the physical soundness of the contemporary active device noise modeling and leading to clear understanding of fluctuation-dissipation relations in nonlinear devices.
\r\n\r\nSecond, we apply statistical electronics to noise problems involved in frequency conversion, an essential function in modern RF and microwave receivers. This study leads to two novel observations of noise figure degradation due to cyclostationary noise and conversion gain enhancement, both dependent on the size of energy storing elements. This novel behavior is experimentally verified with a direct measurement of integrated switching mixers. The results provide new insight into cyclostationary noise processes in frequency conversion and optimum deisgn for switching mixers.
\r\n\r\nThird, application of statistical electronics to noise in frequency generation by self-sustained oscillators leads to a new theory of oscillator noise. This study demonstrates the direct correspondence between the phase noise and the Einstein relation; revealing the underlying physics of oscillator noise. Our approach clarifies the fluctuation-dissipation relation in oscillator noise generation, establishing a link between currently available fluctuation-based and dissipation-based phase noise. models and leading to a clear definition of loaded quality factor of ail oscillator. The novel concepts of virtual damping and linewidth compression put resonators and oscillators in a unified framework, providing immediate design optimization insight. The power of this theoretical development is demonstrated through experimental measurements of various integrated oscillators.
\r\n\r\nOur work on statistical electronics combining circuit engineering and physical science has also resulted in other useful engineering methods, such as graphical optimization, noise simulations for computer-aided design (CAD), and time-varying filter theory.
", "doi": "10.7907/tgja-7y11", "publication_date": "2002", "thesis_type": "phd", "thesis_year": "2002" }, { "id": "thesis:3121", "collection": "thesis", "collection_id": "3121", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-08152001-144501", "primary_object_url": { "basename": "Thesis.pdf", "content": "final", "filesize": 1648934, "license": "other", "mime_type": "application/pdf", "url": "/3121/1/Thesis.pdf", "version": "v2.0.0" }, "type": "thesis", "title": "Fault-tolerant cluster of networking elements", "author": [ { "family_name": "Fan", "given_name": "Chenggong Charles", "clpid": "Fan-C-C" } ], "thesis_advisor": [ { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" } ], "thesis_committee": [ { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "DeHon", "given_name": "Andre", "clpid": "DeHon-A" }, { "family_name": "Effros", "given_name": "Michelle", "clpid": "Effros-M" }, { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" }, { "family_name": "Low", "given_name": "Steven H.", "clpid": "Low-S-H" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "The explosive growth of the Internet demands higher reliability and performance than what the current networking infrastructure can provide. This dissertation explores novel architectures and protocols that provide a methodology for grouping together multiple networking elements, such as routers, gateways, and switches, to create a more reliable and performant distributed networking system. Clustering of networking elements is a novel concept that requires the invention of distributed computing protocols that facilitate efficient and robust support of networking protocols. We introduce the Raincore protocol architecture that achieves these goals by bridging the fields of computer networks and distributed systems.\nIn designing Raincore, we paid special attention to the unique requirements from the networking environment. First, networking clusters need to scale up the networking throughput in addition to the scaling up of computing power. Second, task switching between the different services supported by a networking element has a major negative impact on performance. Third, fast fail-over time is critical for maintaining network connections in the event of failures. We discuss in depth the design of Raincore Group Communication Manager that addresses the forgoing requirements and provides group membership management and reliable multicast transport. It is based on a novel token-ring protocol. We prove that this protocol is formally correct, namely, it satisfies the set of formal specifications that defines the Group Membership problem. \nThe creation of Raincore has already made a substantial impact both at Caltech and the academic community as well as in the industry. The first application is SNOW, a scalable web server cluster that is part of RAIN, a collaborative project between Caltech and JPL/NASA. The second application is RainWall, a commercial solution created by Rainfinity, a Caltech spin-off company, that provides the first fault-tolerant and scalable firewall cluster. These applications exhibit the fast fail-over response, low overhead, and near-linear scalability of the Raincore protocols.\nIn addition, we studied fault-tolerant networking architectures. In particular, we considered efficient constructions of extra-stage fault-tolerant Multistage Interconnection Networks. Multistage Interconnection Networks provide a way to construct a larger switching network using smaller switching elements. We discovered an optimal family of constructions, in the sense that it requires the least number of extra components to tolerate multiple switching element failures. We prove that this is the only family of constructions that has this optimal fault-tolerance property. \n", "doi": "10.7907/R15B-VD58", "publication_date": "2001", "thesis_type": "phd", "thesis_year": "2001" }, { "id": "thesis:3191", "collection": "thesis", "collection_id": "3191", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-08222001-151244", "primary_object_url": { "basename": "HJinthesis.pdf", "content": "final", "filesize": 1146411, "license": "other", "mime_type": "application/pdf", "url": "/3191/1/HJinthesis.pdf", "version": "v2.0.0" }, "type": "thesis", "title": "Analysis and design of turbo-like codes", "author": [ { "family_name": "Jin", "given_name": "Hui", "clpid": "Jin-Hui" } ], "thesis_advisor": [ { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" } ], "thesis_committee": [ { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" }, { "family_name": "Divsalar", "given_name": "Dariush", "clpid": "Divsalar-D" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Effros", "given_name": "Michelle", "clpid": "Effros-M" }, { "family_name": "Dolinar", "given_name": "Samuel J.", "clpid": "Dolinar-S-J" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "50 years after Shannon determined the capacity of memoryless channels, we finally know of practical encoding and decoding algorithms that closely approach this limit. This remarkable feat was first achieved by the invention of turbo codes by Berrou, Glavieux, and Thitimajshima in 1993 cite{BGT93}. Since then turbo codes have essentially revolutionized the coding field and became one of the central research problems in recent years. While there has been a great deal of excellent theoretical work on turbo codes, it is fair to say practice still leads theory by a considerable margin.\r\n \r\nThis thesis endeavors to fill some of that gap. The main body of the thesis concerns coding theorems for general turbo codes. We first prove coding theorems for some simple (yet interesting) serial turbo code ensembles on the AWGN channel. Then we generalize the results for a broader class of turbo-like codes on any memoryless channel. To closely estimate the noise threshold when calculate is feasible, we develop a method based on ``typical pairs decoding'. This method is powerful enough to reproduce Shannon's original coding theorems on any memoryless binary-input symmetric channels. Last we introduce a linearly encodable and decodable code, after carefule design, with performance provably close to Shannon's limit.\r\n \r\nOne main contribution here of both theoretical and practical interests is the introduction of \"repeat-accumulate\" (RA) codes. RA codes are structurally simple enough that a rigorous theoretical analysis is possible, yet random enough that powerful performance under iterative decoding can be obtained. The generalization, irregular RA codes, are shown to be a serious competitor against turbo codes and LDPC codes.", "doi": "10.7907/VC95-2C74", "publication_date": "2001", "thesis_type": "phd", "thesis_year": "2001" }, { "id": "thesis:3236", "collection": "thesis", "collection_id": "3236", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-08272001-155016", "primary_object_url": { "basename": "00ch0.pdf", "content": "final", "filesize": 144139, "license": "other", "mime_type": "application/pdf", "url": "/3236/1/00ch0.pdf", "version": "v2.0.0" }, "type": "thesis", "title": "Why multicast protocols (don't) scale: an analysis of multipoint algorithms for scalable group communication", "author": [ { "family_name": "Schooler", "given_name": "Eve Meryl", "clpid": "Schooler-Eve-Meryl" } ], "thesis_advisor": [ { "family_name": "Chandy", "given_name": "K. Mani", "clpid": "Chandy-K-M" } ], "thesis_committee": [ { "family_name": "Chandy", "given_name": "K. Mani", "clpid": "Chandy-K-M" }, { "family_name": "Martin", "given_name": "Alain J.", "clpid": "Martin-A-J" }, { "family_name": "Estrin", "given_name": "Deborah", "clpid": "Estrin-D" }, { "family_name": "Hickey", "given_name": "Jason J.", "clpid": "Hickey-J-J" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "With the exponential growth of the Internet, there is a critical need to design efficient, scalable and robust protocols to support the network infrastructure. A new class of protocols has emerged to address these challenges, and these protocols rely on a few key techniques, or micro-algorithms, to achieve scalability. By scalability, we mean the ability of groups of communicating processes to grow very large in size. We study the behavior of several of these fundamental techniques that appear in many deployed and emerging Internet standards: Suppression, Announce-Listen, and Leader Election.\r\n\r\nThese algorithms are based on the principle of efficient multipoint communication, often in combination with periodic messaging. We assume a loosely-coupled communication model, where acknowledged messaging among groups of processes is not required. Thus, processes infer information from the periodic receipt or loss of messages from other processes.\r\n\r\nWe present an analysis, validated by simulation, of the performance tradeoffs of each of these techniques. Toward this end, we derive a series of performance metrics that help us to evaluate these algorithms under lossy conditions: expected response time, network usage, memory overhead, consistency attainable, and convergence time. In addition, we study the impact of both correlated and uncorrelated loss on groups of communicating processes.\r\n\r\nAs a result, this thesis provides insights into the scalability of multicast protocols that rely upon these techniques. We provide a systematic framework for calibrating as well as predicting protocol behavior over a range of operating conditions. In the process, we establish a general methodology for the analysis of these and other scalability techniques. Finally, we explore a theory of composition; if we understand the behavior of these micro-algorithms, then we can bound analytically the performance of the more complex algorithms that rely upon them.", "doi": "10.7907/44QZ-R465", "publication_date": "2001", "thesis_type": "phd", "thesis_year": "2001" }, { "id": "thesis:3858", "collection": "thesis", "collection_id": "3858", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-10022001-201911", "primary_object_url": { "basename": "thesis.pdf", "content": "final", "filesize": 827364, "license": "other", "mime_type": "application/pdf", "url": "/3858/1/thesis.pdf", "version": "v2.0.0" }, "type": "thesis", "title": "Analog Computation and Learning in VLSI", "author": [ { "family_name": "Koosh", "given_name": "Vincent Frank", "clpid": "Koosh-Vincent-Frank" } ], "thesis_advisor": [ { "family_name": "Goodman", "given_name": "Rodney M.", "clpid": "Goodman-R-M" } ], "thesis_committee": [ { "family_name": "Goodman", "given_name": "Rodney M.", "clpid": "Goodman-R-M" }, { "family_name": "Martin", "given_name": "Alain J.", "clpid": "Martin-A-J" }, { "family_name": "Diorio", "given_name": "Christopher J.", "clpid": "Diorio-Christopher-J" }, { "family_name": "Koch", "given_name": "Christof", "orcid": "0000-0001-6482-8067", "clpid": "Koch-C" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "Nature has evolved highly advanced systems capable of performing complex computations, adaptation, and learning using analog components. Although digital systems have significantly surpassed analog systems in terms of performing precise, high speed, mathematical computations, digital systems cannot outperform analog systems in terms of power. Furthermore, nature has evolved techniques to deal with imprecise analog components by using redundancy and massive connectivity. In this thesis, analog VLSI circuits are presented for performing arithmetic functions and for implementing neural networks. These circuits draw on the power of the analog building blocks to perform low power and parallel computations.
\r\n\r\nThe arithmetic function circuits presented are based on MOS transistors operating in the subthreshold region with capacitive dividers as inputs to the gates. Because the inputs to the gates of the transistors are floating, digital switches are used to dynamically reset the charges on the floating gates to perform the computations. Circuits for performing squaring, square root, and multiplication/division are shown. A circuit that performs a vector normalization, based on cascading the preceding circuits, is shown to display the ease with which simpler circuits may be combined to obtain more complicated functions. Test results are shown for all of the circuits.
\r\n\r\nTwo feedforward neural network implementations are also presented. The first uses analog synapses and neurons with a digital serial weight bus. The chip is trained in loop with the computer performing control and weight updates. By training with the chip in the loop, it is possible to learn around circuit offsets. The second neural network also uses a computer for the global control operations, but all of the local operations are performed on chip. The weights are implemented digitally, and counters are used to adjust them. A parallel perturbative weight update algorithm is used. The chip uses multiple, locally generated, pseudorandom bit streams to perturb all of the weights in parallel. If the perturbation causes the error function to decrease, the weight change is kept, otherwise it is discarded. Test results are shown of both networks successfully learning digital functions such as AND and XOR.
", "doi": "10.7907/9B65-TB43", "publication_date": "2001", "thesis_type": "phd", "thesis_year": "2001" }, { "id": "thesis:3858", "collection": "thesis", "collection_id": "3858", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-10022001-201911", "primary_object_url": { "basename": "thesis.pdf", "content": "final", "filesize": 827364, "license": "other", "mime_type": "application/pdf", "url": "/3858/1/thesis.pdf", "version": "v2.0.0" }, "type": "thesis", "title": "Analog Computation and Learning in VLSI", "author": [ { "family_name": "Koosh", "given_name": "Vincent Frank", "clpid": "Koosh-Vincent-Frank" } ], "thesis_advisor": [ { "family_name": "Goodman", "given_name": "Rodney M.", "clpid": "Goodman-R-M" } ], "thesis_committee": [ { "family_name": "Goodman", "given_name": "Rodney M.", "clpid": "Goodman-R-M" }, { "family_name": "Martin", "given_name": "Alain J.", "clpid": "Martin-A-J" }, { "family_name": "Diorio", "given_name": "Christopher J.", "clpid": "Diorio-Christopher-J" }, { "family_name": "Koch", "given_name": "Christof", "orcid": "0000-0001-6482-8067", "clpid": "Koch-C" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "Nature has evolved highly advanced systems capable of performing complex computations, adaptation, and learning using analog components. Although digital systems have significantly surpassed analog systems in terms of performing precise, high speed, mathematical computations, digital systems cannot outperform analog systems in terms of power. Furthermore, nature has evolved techniques to deal with imprecise analog components by using redundancy and massive connectivity. In this thesis, analog VLSI circuits are presented for performing arithmetic functions and for implementing neural networks. These circuits draw on the power of the analog building blocks to perform low power and parallel computations.
\r\n\r\nThe arithmetic function circuits presented are based on MOS transistors operating in the subthreshold region with capacitive dividers as inputs to the gates. Because the inputs to the gates of the transistors are floating, digital switches are used to dynamically reset the charges on the floating gates to perform the computations. Circuits for performing squaring, square root, and multiplication/division are shown. A circuit that performs a vector normalization, based on cascading the preceding circuits, is shown to display the ease with which simpler circuits may be combined to obtain more complicated functions. Test results are shown for all of the circuits.
\r\n\r\nTwo feedforward neural network implementations are also presented. The first uses analog synapses and neurons with a digital serial weight bus. The chip is trained in loop with the computer performing control and weight updates. By training with the chip in the loop, it is possible to learn around circuit offsets. The second neural network also uses a computer for the global control operations, but all of the local operations are performed on chip. The weights are implemented digitally, and counters are used to adjust them. A parallel perturbative weight update algorithm is used. The chip uses multiple, locally generated, pseudorandom bit streams to perturb all of the weights in parallel. If the perturbation causes the error function to decrease, the weight change is kept, otherwise it is discarded. Test results are shown of both networks successfully learning digital functions such as AND and XOR.
", "doi": "10.7907/9B65-TB43", "publication_date": "2001", "thesis_type": "phd", "thesis_year": "2001" }, { "id": "thesis:4371", "collection": "thesis", "collection_id": "4371", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-11022006-104150", "primary_object_url": { "basename": "Goncalves_l_2000.pdf", "content": "final", "filesize": 12342869, "license": "other", "mime_type": "application/pdf", "url": "/4371/1/Goncalves_l_2000.pdf", "version": "v3.0.0" }, "type": "thesis", "title": "Automatic observation and synthesis of human motion", "author": [ { "family_name": "Goncalves", "given_name": "Luis", "clpid": "Goncalves-L" } ], "thesis_advisor": [ { "family_name": "Perona", "given_name": "Pietro", "clpid": "Perona-P" } ], "thesis_committee": [ { "family_name": "Perona", "given_name": "Pietro", "clpid": "Perona-P" }, { "family_name": "Barr", "given_name": "Alan H.", "clpid": "Barr-A-H" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Burdick", "given_name": "Joel Wakeman", "clpid": "Burdick-J-W" }, { "family_name": "Andersen", "given_name": "Richard A.", "clpid": "Andersen-R-A" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "Over the past few decades Computer Vision and Computer Graphics have experienced a rapid evolution, thanks in part to the continual improvement in computer hardware, which enables the investigation of increasingly complex problems.\r\n\r\nIn Computer Graphics this evolution is visible on a nearly day-by-day basis. For instance, computer-generated special effects in feature films have evolved to such a level of sophistication that it is often impossible to distinguish what is real from what is not. However, one challenging problem that still stands, considered by many experts in the field to be a Holy-Grail of Computer Graphics, is the automatic synthesis of life-like human character animation. Although rendering and modeling techniques have reached a stage where a computer generated image of a person is nearly indistinguishable from the real thing, as soon as that model begins to move the illusion is broken. The problem is difficult because no-one yet knows how to model human motion in all it's intricacy and subtlety, and also because humans are so well tuned to perceive these subtleties that they can only be fooled if the modeling is done with complete perfection.\r\n\r\nIn this thesis, we explore a novel method of automatic synthesis of human motion that brings us one step closer to the ultimate goal. The method is based on decomposing human motion into elemental, nameable actions such as walking, running, and throwing, and using observations of people performing these actions to create mathematical models of the actions. Various samples of an action are acquired, and each sample is labeled according to state (initial body configuration), goal (desired outcome of the motion, such as direction of a throw or placement of a foot for a step), and mood and style parameters. Then established and novel techniques of machine learning are applied to derive a function that can synthesis a motion given some desired parameters. We explore the use of polynomial interpolants, radial basis function networks (RBFs), feed-forward neural networks (FFNNs) with sigmoidal activation functions, as well as a new method with local linear models. We find that a linear model more often that not works quite well, whereas higher order polynomial interpolants, RBFs and FFNNs are unable to extrapolate robustly when the motion parameters lie outside of the convex hull of the parameters of the available sample motions. The method with local linear models successfully improves the fidelity of the synthetic motions compared to the linear model, and also provides robust extrapolation. We also investigate the use of a recursive, probabilitic model where motions are specified by defining the initial and final body poses of the motion, and synthesis is done by computing the most likely motion to satisfy the boundary constraints. Although the results with this method are not yet completely satisfactory, it holds promise, and under certain types of conditions can re-synthesize the sample motions more accurately than any of the other methods.\r\n\r\nWith the additional development of methods to smoothly concatenate actions together and to interactively map synthesized motions to a 3-D polygonal character model, a realtime interactive demo was created that successfully demonstrates the level of realism and interactivity achievable by our method of human motion synthesis.\r\n\r\nOur interest in the problem of realistic human motion synthesis arose from an initial study of the (in some sense) inverse problem in Computer Vision of the automatic observation (rather than synthesis) of human motion. Although progress in Computer Vision has not yet reached a level enabling it's widespread use in daily life, this state will most likely be achieved within the next decade. One large class of problems for which this is the case is the endowment of computers with visual perceptual skills similar to those of humans. Among the vast set of visual tasks imagineable, the automatic detection, recognition, and estimation of humans and human motion is a particularly interesting set of problems since there are many possible applications of such a technology in modern life, ranging from security and monitoring systems, to systems for biometric analysis, to novel human-machine interfaces.\r\n\r\nIn this thesis we describe a method of robustly estimating the motion of a human body from a monocular view. The method is based on the use of a 3-D model of the body, and comparing the actual image to an expected image based on the 3-D model to update the estimate of the body pose at each time step. The method was implemented in realtime as a human-machine interface. This system demonstrated that the method can be used to robustly track a human arm with a hand-tip positioning resolution of 2cm under close viewing conditions (where perspective projection causes significant changes in the appearance of the arm in the camera view).\r\n", "doi": "10.7907/n3fn-jd79", "publication_date": "2000", "thesis_type": "phd", "thesis_year": "2000" }, { "id": "thesis:687", "collection": "thesis", "collection_id": "687", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-02202008-104935", "primary_object_url": { "basename": "Kirac_a_1999.pdf", "content": "final", "filesize": 6313690, "license": "other", "mime_type": "application/pdf", "url": "/687/1/Kirac_a_1999.pdf", "version": "v3.0.0" }, "type": "thesis", "title": "Optimal orthonormal subband coding and lattice quantization with vector dithering", "author": [ { "family_name": "Kirac", "given_name": "Ahmet", "clpid": "Kirac-A" } ], "thesis_advisor": [ { "family_name": "Vaidyanathan", "given_name": "P. P.", "clpid": "Vaidyanathan-P-P" } ], "thesis_committee": [ { "family_name": "Vaidyanathan", "given_name": "P. P.", "clpid": "Vaidyanathan-P-P" }, { "family_name": "Fadavi", "given_name": "Jalil", "clpid": "Fadavi-J" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Simon", "given_name": "Marvin K.", "clpid": "Simon-M-K" }, { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "In the digital era that we live in, efficient coding of signals is an unquestionable need. This thesis is about one of the most useful and popular technique of digital coding: subband coding. Subband coding and its cousin wavelet-based coding are now the preferred methods for not only speech, but also audio, image, and video signals. Subband coding involves a linear part which is a filter bank, and a nonlinear part which is usually a uniform scalar quantization of each of the subbands. Subband coders are classified according to the type of filter bank used for its transform. This thesis is mainly about orthonormal subband coding. The ability of an orthonormal filter bank to decompose the signal into components that have a diverse set of signal energies is an indicator of its efficiency for subband coding. Such a diversity in the set of the subband energies is fully utilized by a process called bit allocation. The traditional results on the optimality of a filter bank for given input statistics assume that the quantizers operate at high bit rates.\n\nThis thesis presents optimality results under more general quantizer models without assuming high bit rates. This is accomplished by revealing the relationship between the problems of optimal orthonormal subband coding and principal component representation of signals. The latter is done using what is called a principal component filter bank (PCFB). A PCFB is one that compacts most of the energy of a signal into smaller subsets of subbands. To date, there has not been significant theoretical developments in the field of optimal nonuniform subband coding, although the successful techniques of wavelet-based coding are among the state of the art in practice. Such techniques utilize a form of a nonuniform filter bank with a certain structure which makes it efficient for its implementation. In this thesis, we provide optimality results for the nonuniform orthonormal subband coding as well. As in the uniform case, the principal component representation of signals continues to play the key role. We introduce nonuniform PCFB's and link them to the optimal subband coding problem. A PCFB, in particular, contains a filter that compacts most of the signal energy into one single channel: energy compaction filter. The thesis goes into details of designing such filters optimally. In particular, we propose an analytical method in the two-channel case and a very efficient window method in the arbitrary M\u2014channel case. Multistage design of compaction filters has also been worked out.\n\nFinally we extend the analysis of uniform scalar quantization to multiple dimensions. We provide an exact statistical relationship between a lattice quantizer noise and its input vector. We then extend the idea of dithering to the vector case. Dithering is a means of statistically rendering the quantization noise independent of the input. We address the optimal choice of a lattice for a given dimension and also optimal pre- and post-filtering of a dithered lattice quantizer.\n", "doi": "10.7907/ec82-t391", "publication_date": "1999", "thesis_type": "phd", "thesis_year": "1999" }, { "id": "thesis:771", "collection": "thesis", "collection_id": "771", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-02262008-093428", "primary_object_url": { "basename": "Xu_m_1999.pdf", "content": "final", "filesize": 3766240, "license": "other", "mime_type": "application/pdf", "url": "/771/1/Xu_m_1999.pdf", "version": "v3.0.0" }, "type": "thesis", "title": "Iterative decoding and graphical code representations", "author": [ { "family_name": "Xu", "given_name": "Meina", "clpid": "Xu-Meina" } ], "thesis_advisor": [ { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" } ], "thesis_committee": [ { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" }, { "family_name": "Divsalar", "given_name": "Dariush", "clpid": "Divsalar-D" }, { "family_name": "Kaleh", "given_name": "Ghassan Kawas", "clpid": "Kaleh-G-K" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Franklin", "given_name": "Joel N.", "clpid": "Franklin-J-N" }, { "family_name": "Tanner", "given_name": "Michael", "clpid": "Tanner-M" }, { "family_name": "Vaidyanathan", "given_name": "P. P.", "clpid": "Vaidyanathan-P-P" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "Since the invention of turbo codes, there has been an explosion of interest in iterative decoding and graphical representation of codes. This thesis examines the iterative decoding of codes defined on graphs with cycles, which appears to be an efficient means of achieving the Shannon limit. Much of this analysis is on the iterative min-sum decoding of tail-biting codes and cycle codes. We have identified the pseudocodeword as the cause of the suboptimal performance of the iterative decoder, and we have obtained a union bound for the performance of the iterative decoder on both AWGN and BSC channels. Using the union bound argument, for cycle codes, we have shown that the performance of the iterative decoder is asymptotically as good as that of the ML decoder. As for tail-biting codes, the same thing is true if the lowest weight pseudocodeword is at least the minimum weight of the code. Unfortunately, the analysis of tail-biting codes and cycle codes does not extend to turbo codes and low density parity check codes in general. Our next approach is to determine the average behavior of message passing algorithms by studying the evolution of their \"message\" densities. For the class of \"repeat and accumulate\" serially concatenated turbo-like codes, we have devised an algorithm for determining their \"threshold\" values. When the signal-to-noise ratio is larger than the threshold value, the error probabilities of message passing algorithms approach zero, whereas if the signal-to-noise ratio is less than the threshold value, the error probability stays bounded away from zero. Some message passing algorithms and graphical representation of codes are efficient means of devising ML or MAP decoding algorithms. We have proposed a junction tree representation for linear block codes, and we have shown that the minimum junction tree can be less complex than the minimal trellis.", "doi": "10.7907/q731-9q50", "publication_date": "1999", "thesis_type": "phd", "thesis_year": "1999" }, { "id": "thesis:1834", "collection": "thesis", "collection_id": "1834", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-05162005-084223", "primary_object_url": { "basename": "00_cover.pdf", "content": "final", "filesize": 7260, "license": "other", "mime_type": "application/pdf", "url": "/1834/1/00_cover.pdf", "version": "v2.0.0" }, "type": "thesis", "title": "Highly available distributed storage systems", "author": [ { "family_name": "Xu", "given_name": "Lihao", "clpid": "Xu-Lihao" } ], "thesis_advisor": [ { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" } ], "thesis_committee": [ { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "van Tilborg", "given_name": "Henk C.A.", "clpid": "van-Tilborg-H-C-A" }, { "family_name": "Chandy", "given_name": "K. Mani", "clpid": "Chandy-K-M" }, { "family_name": "Vaidyanathan", "given_name": "P. P.", "clpid": "Vaidyanathan-P-P" }, { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "As the need for data explodes with the passage of time and the increase of computing power, data storage becomes more and more important. Distributed storage, as distributed computing before it, is coming of age as a good solution to make systems highly available, i.e., highly scalable, reliable and efficient. The focus of this thesis is how to achieve data reliability and efficiency in distributed storage systems. This thesis consists of two parts. The first part deals with the reliability of distributed storage systems. Reliability is achieved by computationally efficient MDS array codes that eliminate single points of failure in the systems, thus providing more reliability and flexibility to the systems. Such codes can be used as general MDS error-correcting codes. They are particularly suitable for use in distributed storage systems. The second part deals with the efficiency of distributed storage systems. Methods are proposed to improve the performance of data server and storage systems significantly through the proper use of data redundancy. These methods are based on error-correcting codes, particularly the MDS array codes developed in the first part.\r\n\r\nTwo new classes of MDS array codes are presented: the X-Code and the B-Code. The encoding operations of both codes are optimal, i.e., their update complexity achieves the theoretical lower bound. They distribute parity bits over all columns rather than concentrating them on some parity columns. As with other array codes, the error model for both codes is that errors or erasures are columns of the array, i.e., if at least one bit of a column is an error or erasure, then the whole column is considered to be an error or erasure. Both codes are of distance 3, i.e., they can either: correct two erasures, detect two errors or correct one error. In addition to encoding algorithms, efficient decoding algorithms are proposed, both for erasure-correcting and for error-correcting. In fact, the erasure-correcting algorithms are also optimal in terms of computation complexity.\r\n\r\nThe X-Code has a very simple geometrical structure: the parity bits are constructed along two groups of parallel parity lines of slopes 1 and -1. This is the origin of the name X-Code. This simple geometrical structure allows simple erasure-decoding and error-decoding algorithms, using only XORs and vector cyclic-shift operations.\r\n\r\nThe significance of the B-code not only includes all its optimality properties: MDS, optimal encoding and optimal decoding, but also its relation with a 3-decade old graph theory problem. It is proven in this thesis that constructing a B-Code of odd length is exactly equivalent to constructing a perfect one-factorization (or P1F) of a complete graph. Constructing a P1F of an arbitrary complete graph has remained a conjecture since the early 1960's. Though the P1F conjecture remains unsolved, the B-code as the first real application of the P1F problem will hopefully spur more research on it. It is also conjectured in this thesis that constructing a B-Code of any length, even or odd, is equivalent to constructing a P1F of a complete graph. An efficient error-correcting algorithm for the B-Code is also presented, which is based on the relations between the B-Code and its dual. The algorithm might give a hint of how to develop efficient decoding algorithms for other codes.\r\n\r\nWhile it is intuitive that redundancy can bring reliability to a system, this thesis gives another direction: using redundancy actively to improve performance (efficiency) of distributed data systems. The results in this direction are both theoretical and experimental. System models are extracted from experiments in real practical systems; analytical results are derived using these and are then fed back to experiments for verification.\r\n\r\nIn this thesis, a novel deterministic voting scheme that uses error-correcting codes is proposed. The voting scheme generalizes all known simple deterministic voting algorithms. It can be tuned to various application environments with different error rates to drastically reduce average communication complexity, i.e., the amount of information that must be transmitted in order to get correct voting results.\r\n\r\nTwo problems are identified to improve the performance of general data server systems, namely the data distribution problem and the data acquisition problem. Solutions to these are proposed, as are general analytical results on performance of (n, k) systems. A simple service time model of a practical disk-based distributed server system is given. This model, which is based on experimental results, is a starting point for data distribution and data acquisition schemes. These results, both experimental and analytical, can be further used for more sophisticated scheduling schemes to optimize or improve the performance of data server systems that serve multiple clients simultaneously.\r\n\r\nFinally, some research problems related to storage systems are proposed as future directions.", "doi": "10.7907/EQK9-8C84", "publication_date": "1999", "thesis_type": "phd", "thesis_year": "1999" }, { "id": "thesis:3323", "collection": "thesis", "collection_id": "3323", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-09042007-091804", "primary_object_url": { "basename": "Ofria_ca_1999.pdf", "content": "final", "filesize": 13900717, "license": "other", "mime_type": "application/pdf", "url": "/3323/1/Ofria_ca_1999.pdf", "version": "v3.0.0" }, "type": "thesis", "title": "Evolution of genetic codes", "author": [ { "family_name": "Ofria", "given_name": "Charles A.", "orcid": "0000-0003-2924-1732", "clpid": "Ofria-C-A" } ], "thesis_advisor": [ { "family_name": "Adami", "given_name": "Christoph Carl", "clpid": "Adami-C-C" } ], "thesis_committee": [ { "family_name": "Adami", "given_name": "Christoph Carl", "clpid": "Adami-C-C" }, { "family_name": "Barr", "given_name": "Alan H.", "clpid": "Barr-A-H" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Allman", "given_name": "John M.", "clpid": "Allman-J-M" }, { "family_name": "Abu-Mostafa", "given_name": "Yaser S.", "clpid": "Abu-Mostafa-Y-S" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "In this thesis, I use analytical and computational techniques to study the development of codes in evolutionary systems. We only know of one instance of such a genetic code in the natural world: our own DNA. However, the results from my work are expected to be universally true for all evolving systems. I use mathematical models and conduct experiments with avida, a software-based research platform for the study of evolution in \"digital organisms.\" This allows me to collect statistically powerful data over evolutionary timescales infeasible in a biological system.\r\n\r\nIn the avida system, Darwinian evolution is implemented on populations of self-replicating computer programs. A typical experiment is seeded with a single ancestor program capable only of reproduction. This ancestor gives rise to an entire population of programs, which adapt to interact with a complex environment, while developing entirely new computational capabilities. I study the process of evolution in this system, taking exact measurements on the underlying genetic codes, and performing tests that would be prohibitively difficult in biological systems.\r\n\r\nI have focused on the following areas in studying the evolution of genetic codes: \r\n\r\nInformation Theory: I treat the process of reproduction as a noisy channel in which codes are transmitted from the parent's genome to the child. Unlike most channels, however, evolution actively selects for codes received with a higher information content, even if this increased information was introduced via noise. A genetic code consists of information about the environment surrounding the organism. As a population adapts, this information increases, and can be approximated through measuring the reduction of per-nucleotide entropy - in effect sites freeze in place as they code for useful functionality. In the avida system, we know the sequence of all genomes in the population, and new computational genes can be identified as they are formed.\r\n\r\nThe Evolution of Genetic Organization: Organisms incapable of error correction (such as viruses) develop strong code compaction techniques to minimize their target area for mutations, the most prominent of which is overlapping genes. Higher organisms, however, are capable of reducing their mutational load and will explicitly spread out their code, cleanly segregating their genes. I investigate the pressures behind overlapping or segregation of genes, and demonstrate that overlaps have a side effect of drastically reducing the probability of neutral mutations within a gene, and hence hindering continued adaptation. Further, in a changing environment, overlapping genes have a significantly reduced ability to adapt independently. I compare overlapping and singly expressed sections of code in avida, and show a significant (two-fold) difference in the average per-site variation. I also demonstrate the evolutionary pressure for organisms to segregate their genes in a fluctuating environment to improve their adaptive abilities.\r\n\r\nEvolving Computer Programs: I explore evolution in digital genetic codes, and isolate some of those features of a programming language that promote continuous adaptation. In the biological world evolution gives rise to complex organisms robust to changing situations in their environment. This increase in complexity and \"functionality\" of the organisms typically generates more stable systems. On the other hand, as computer programs gain complexity, they only become more fragile. If two programs interact in a way not explicitly designed, the results are neither predictable nor reliable. In fact, computer programs often fail even when put to the use for which they were explicitly intended. Computational organisms, however, have a level of robustness more akin to their biological counterparts, not only performing computations, but often doing so in a manner beyond the efficiency that a human programmer could produce.\r\n\r\nFinally, all of this work is tied together, and future directions for its continuation are explored.", "doi": "10.7907/2z40-1m97", "publication_date": "1999", "thesis_type": "phd", "thesis_year": "1999" }, { "id": "thesis:3689", "collection": "thesis", "collection_id": "3689", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-09222005-110351", "primary_object_url": { "basename": "Sill_j_1998.pdf", "content": "final", "filesize": 3275897, "license": "other", "mime_type": "application/pdf", "url": "/3689/1/Sill_j_1998.pdf", "version": "v2.0.0" }, "type": "thesis", "title": "Monotonicity and connectedness in learning systems", "author": [ { "family_name": "Sill", "given_name": "Joseph", "clpid": "Sill-J" } ], "thesis_advisor": [ { "family_name": "Abu-Mostafa", "given_name": "Yaser S.", "clpid": "Abu-Mostafa-Y-S" } ], "thesis_committee": [ { "family_name": "Abu-Mostafa", "given_name": "Yaser S.", "clpid": "Abu-Mostafa-Y-S" }, { "family_name": "Psaltis", "given_name": "Demetri", "clpid": "Psaltis-D" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "This thesis studies two properties- monotonicity and connectedness- in the context of machine learning. The first part of the thesis examines the role of monotonicity constraints in machine learning from both practical and theoretical perspectives. Two techniques for enforcing monotonicity in machine learning models are proposed. The first method adds to the objective function a penalty term measuring the degree to which the model violates monotonicity. The penalty term can be interpreted as a Bayesian prior favoring functions which obey monotonicity. This method has the potential to enforce monotonicity only approximately, making it appropriate for situations where strict monotonicity may not hold. The second approach consists of a model which is monotonic by virtue of functional form. This model is shown to have universal approximation capabilities with respect to the class M of monotonic functions. A variety of theoretical results are also presented regarding M. The generalization behavior of this class is shown to depend heavily on the probability distribution over the input space. Although the VC dimension of M is [infinity], the VC entropy (i.e., the expected number of dichotomies) is modest for many distributions, allowing us to obtain bounds on the generalization error. Monte Carlo techniques for estimating the capacity and VC entropy of M are presented.\n\nThe second part of the thesis considers broader issues in learning theory. Generalization error bounds based on the VC dimension describe a function class by counting the number of dichotomies it induces. In this thesis, a more detailed characterization is presented which takes into account the diversity of a set of dichotomies in addition to its cardinality. Many function classes in common usage are shown to possess a property called connectedness. Models with this property induce dichotomy sets which are highly clustered and have little diversity. We derive an improvement to the VC bound which applies to function classes with the connectedness property.\n", "doi": "10.7907/GQWN-1H71", "publication_date": "1998", "thesis_type": "phd", "thesis_year": "1998" }, { "id": "thesis:1775", "collection": "thesis", "collection_id": "1775", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-05132005-143440", "primary_object_url": { "basename": "Bohossian_v_1998.pdf", "content": "final", "filesize": 1050812, "license": "other", "mime_type": "application/pdf", "url": "/1775/1/Bohossian_v_1998.pdf", "version": "v2.0.0" }, "type": "thesis", "title": "Neural logic : theory and implementation", "author": [ { "family_name": "Bohossian", "given_name": "Vasken Z.", "clpid": "Bohossian-Vasken-Z" } ], "thesis_advisor": [ { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" } ], "thesis_committee": [ { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Barr", "given_name": "Alan H.", "clpid": "Barr-A-H" }, { "family_name": "Koch", "given_name": "Christof", "orcid": "0000-0001-6482-8067", "clpid": "Koch-C" }, { "family_name": "Perona", "given_name": "Pietro", "orcid": "0000-0002-7583-5809", "clpid": "Perona-P" }, { "family_name": "Abu-Mostafa", "given_name": "Yaser S.", "clpid": "Abu-Mostafa-Y-S" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "NOTE: Text or symbols not renderable in plain ASCII are indicated by [...]. Abstract is included in .pdf document.\r\n\r\nHuman brains are by far superior to computers in solving hard problems like combinatorial optimization and image and speech recognition, although their basic building blocks are several orders of magnitude slower. This observation has boosted interest in the field of artificial networks [20], [37]. The latter are built by interconnecting artificial neurons whose behavior is inspired by that of biological neurons. In this thesis we consider the Boolean version of an artificial neuron, namely, a Linear Threshold (LT) element, which computes a neural-like Boolean function of n binary inputs [32]. An LT element outputs the sign of a weighted sum of its Boolean inputs. The main issues in the study of networks (circuits) consisting of LT elements, called LT circuits, include the estimation of their computational capabilities and limitations and the comparison of their properties with those of traditional Boolean logic circuits based on AND, OR and NOT gates (called AON circuits). For example, there is a strong evidence that LT circuits are more efficient than AON circuits in implementing a number of important functions including the addition, product and division of integers [44], [45].\r\n\r\nIt is easy to see that an LT element is more powerful than an AON gate, simply because of the freedom one has in selecting the weights. Indeed, different choices of weights produce different Boolean functions. As a matter of fact, the number of n-input Boolean functions that can be implemented by a single LT element is of the order of [\u2026], [42], [22]. That additional power comes at the cost of added complexity. Some LT functions require weights that are very different in magnitude, potentially rendering difficult hardware or software implementations of the corresponding LT elements. For that reason, theoretical research in the field of LT circuits has focused on the weights, in particular the power of LT elements with restricted weights. As early as 1971, Muroga, [32], proved that any linear threshold element can be implemented with integer weights. That is, by restricting the magnitudes of the weights to natural numbers, one does not lose any power of the original LT element. We generalize this result to arbitrary subsets of the set of real numbers. For example, we show that one can restrict the weights to be the square of integers, and still be able to realize all LT functions. We ask the following question. What are the conditions on the subset [\u2026] which guarantee that all LT functions can be implemented with weights drawn from it?\r\n\r\nAnother aspect of the complexity of the weights is their growth as the number of inputs increases. It has been shown [17], [33], [38], [43] that there exist linear threshold functions that can be implemented by a single threshold element with exponentially growing weights, but cannot be implemented by a threshold element with smaller polynomialy growing weights. In light of that result the above question was dealt with by defining a class, called [\u2026], within the set of linear threshold functions: the class of functions with \"small\" (i.e. polynomialy growing) weights [43]. We focus on a single LT element. Our contribution consists in two novel methods for constructing threshold functions with minimal weights, which allow us to fill up the gap between polynomial and exponential weight growth by further refining the separation. Namely, we prove that the class of linear threshold functions with polynomial-size weights can be divided into subclasses [\u2026], according to the degree, d, of the polynomial. In fact, we prove a more general result-\u2014that there exists a linear threshold function for any arbitrary number of inputs and any weight size.\r\n\r\nEven though some LT functions require weights that grow exponentially with the number of input variables, it has been shown recently, in [13], [18], that such functions can be replaced by a two-layer circuit composed of LT gates with polynomially growing, i.e., small weights. We improve the best known bound on the size of that circuit, presented in [18] by focusing on a particular function with large coefficients. We also derive explicit two-layer circuits. Two layer LT circuits are in general composed of different linear threshold elements, but for some useful Boolean functions, such as parity, addition and product, the gates of the first layer are almost identical. To take advantage of this fact we introduce a new Boolean computing element. Instead of the sign function, it computes an arbitrary (with polynomialy many transitions) Boolean function of the weighted sum of its inputs. We call the new computing element an LTM element, which stands for Linear Threshold with Multiple transitions. The advantages of LTM become apparent in the context of VLSI implementation. Indeed, this new model reduces the layout area of the corresponding symmetric function from [\u2026] to O(n). We present a VLSI implementations of both LT and LTM elements. Two kinds of elements were fabricated, programmable and hardwired. The programmable elements use the charge on a floating gate in order to store the values of the weights.\r\n\r\nFor many years, the topic of linear threshold logic, has been approached in two different ways, theory, i.e. computational circuit complexity, [38], [56], and hardware implementation, [48], [40]. Surprisingly, there has been very little interaction between those two approaches. As a whole, the present thesis is one step towards establishing a connection between the theory and implementation of threshold circuits. Its contributions are at three levels. At the theoretical level, new classes of functions such as [\u2026] and LTM are defined and their computational power is estimated. At the algorithmic level, we show how to convert real weights to weights drawn from arbitrary subset of the real numbers, e.g., integer weights, we also show how to construct LT functions with minimal weights, and finally we present an algorithm that produces an [\u2026] circuit (circuit composed of gates with small weights), that computes the comparison function, COMP. We also present LTM circuits computing useful functions, such as XOR, ADD, PRODUCT. At the implementation level, we show the design, layout and testing of the VLSI implementation of LT and LTM. Establishing a connection between the theoretical and practical aspects of threshold logic will profit both domains by providing solutions for practical problems and by defining new theoretical questions inspired by implementation issues.\r\n", "doi": "10.7907/N0T5-7J92", "publication_date": "1998", "thesis_type": "phd", "thesis_year": "1998" }, { "id": "thesis:1013", "collection": "thesis", "collection_id": "1013", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-03192008-090303", "primary_object_url": { "basename": "Yu_z_1997.pdf", "content": "final", "filesize": 4659562, "license": "other", "mime_type": "application/pdf", "url": "/1013/1/Yu_z_1997.pdf", "version": "v3.0.0" }, "type": "thesis", "title": "Design Issues in Communications Networks: Reliability and Traffic Analysis", "author": [ { "family_name": "Yu", "given_name": "Zhong", "clpid": "Yu-Zhong" } ], "thesis_advisor": [ { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" } ], "thesis_committee": [ { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" }, { "family_name": "Goldsmith", "given_name": "Andrea Jo", "orcid": "0000-0001-5686-800X", "clpid": "Goldsmith-A-J" }, { "family_name": "Posner", "given_name": "Edward C.", "clpid": "Posner-E-C" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Vaidyanathan", "given_name": "P. P.", "orcid": "0000-0003-3003-7042", "clpid": "Vaidyanathan-P-P" }, { "family_name": "Wilson", "given_name": "Richard M.", "clpid": "Wilson-R-M" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "This thesis aims to investigate two rather separate issues: network reliability and traffic analysis. The first concerns the reliability for unreliable systems, including communications networks with possible link failures, and more general fault-tolerant systems. The second concerns the traffic characteristics specifically in ATM networks with respect to the performance of statistical multiplexers.\r\n\r\nOne way in which we studied the reliability issue is via mean time to failure (MTTF) which considers systems that have component failures and repairs with exponential distributions. Such systems can be modeled by continuous-time discrete- state Markov chains. We investigated the MTTF from a more general framework of fault-tolerant systems (FTS), and developed two systematic approaches, the allpath-weight approach and the signal-flow-graph approach, to compute the MTTF. We also derived a simple asymptotic formula for estimating the MTTF, and obtained asymptotically the optimal networks in terms of the MTTF.\r\n\r\nThe other way in which we studied the reliability issue is via reliability polynomials for a system with component failures with certain fixed probability that is independent of time, but a function of the size of the system. No repair is allowed. We modeled such systems by random graphs, and analyzed reliability polynomials in a framework of random graph theory. We specifically focused on certain regular random graphs and analyzed the evolution of the regular random graphs, by showing a transition phenomenon when such a regular random graph evolves from edge probability zero to probability one because of the expansion of graph size, and identified its threshold function. Our work extends the study of the evolution of random graphs to regular random graphs which do not appear in the literature of random graphs, and our results are generalizations of some famous previously known results in random graph theory.\r\n\r\nAs for the second issue of traffic analysis in ATM networks, we first studied, via the approach of generating functions, Markov on-off traffic and the performance behavior of a statistical multiplexer with such traffic. We developed a heuristic procedure which allowed us to compute the expected buffer occupancy of a statistical multiplexer with Markov on-off traffic, and obtained closed form formulas showing that the expected buffer occupancy under such traffic not only depends on the incoming traffic intensity, but also largely on the burstiness of incoming traffic. The expected buffer occupancy becomes unbounded with large enough traffic burstiness, even though the traffic intensity is small. These results showed that burstiness control of traffic was very critical in designing ATM networks.\r\n\r\nWe then introduced a class of burst-constrained traffic sources, the periodic interchangeable (PI) traffic, and applied generalized Ballot theorems to analyze the buffer occupancy in a statistical multiplexer with PI traffic. We derived closed form formulas for survivor functions, expected buffer occupancy, and simple asymptotic formula that can be used as a rule of thumb for dimensioning buffer size in designing a statistical multiplexer. The results obtained could shed light on the study of worst case performance of statistical multiplexers for burst-constrained traffic sources in ATM networks.", "doi": "10.7907/pa8m-ay72", "publication_date": "1997", "thesis_type": "phd", "thesis_year": "1997" }, { "id": "thesis:1013", "collection": "thesis", "collection_id": "1013", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-03192008-090303", "primary_object_url": { "basename": "Yu_z_1997.pdf", "content": "final", "filesize": 4659562, "license": "other", "mime_type": "application/pdf", "url": "/1013/1/Yu_z_1997.pdf", "version": "v3.0.0" }, "type": "thesis", "title": "Design Issues in Communications Networks: Reliability and Traffic Analysis", "author": [ { "family_name": "Yu", "given_name": "Zhong", "clpid": "Yu-Zhong" } ], "thesis_advisor": [ { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" } ], "thesis_committee": [ { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" }, { "family_name": "Goldsmith", "given_name": "Andrea Jo", "orcid": "0000-0001-5686-800X", "clpid": "Goldsmith-A-J" }, { "family_name": "Posner", "given_name": "Edward C.", "clpid": "Posner-E-C" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Vaidyanathan", "given_name": "P. P.", "orcid": "0000-0003-3003-7042", "clpid": "Vaidyanathan-P-P" }, { "family_name": "Wilson", "given_name": "Richard M.", "clpid": "Wilson-R-M" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "This thesis aims to investigate two rather separate issues: network reliability and traffic analysis. The first concerns the reliability for unreliable systems, including communications networks with possible link failures, and more general fault-tolerant systems. The second concerns the traffic characteristics specifically in ATM networks with respect to the performance of statistical multiplexers.\r\n\r\nOne way in which we studied the reliability issue is via mean time to failure (MTTF) which considers systems that have component failures and repairs with exponential distributions. Such systems can be modeled by continuous-time discrete- state Markov chains. We investigated the MTTF from a more general framework of fault-tolerant systems (FTS), and developed two systematic approaches, the allpath-weight approach and the signal-flow-graph approach, to compute the MTTF. We also derived a simple asymptotic formula for estimating the MTTF, and obtained asymptotically the optimal networks in terms of the MTTF.\r\n\r\nThe other way in which we studied the reliability issue is via reliability polynomials for a system with component failures with certain fixed probability that is independent of time, but a function of the size of the system. No repair is allowed. We modeled such systems by random graphs, and analyzed reliability polynomials in a framework of random graph theory. We specifically focused on certain regular random graphs and analyzed the evolution of the regular random graphs, by showing a transition phenomenon when such a regular random graph evolves from edge probability zero to probability one because of the expansion of graph size, and identified its threshold function. Our work extends the study of the evolution of random graphs to regular random graphs which do not appear in the literature of random graphs, and our results are generalizations of some famous previously known results in random graph theory.\r\n\r\nAs for the second issue of traffic analysis in ATM networks, we first studied, via the approach of generating functions, Markov on-off traffic and the performance behavior of a statistical multiplexer with such traffic. We developed a heuristic procedure which allowed us to compute the expected buffer occupancy of a statistical multiplexer with Markov on-off traffic, and obtained closed form formulas showing that the expected buffer occupancy under such traffic not only depends on the incoming traffic intensity, but also largely on the burstiness of incoming traffic. The expected buffer occupancy becomes unbounded with large enough traffic burstiness, even though the traffic intensity is small. These results showed that burstiness control of traffic was very critical in designing ATM networks.\r\n\r\nWe then introduced a class of burst-constrained traffic sources, the periodic interchangeable (PI) traffic, and applied generalized Ballot theorems to analyze the buffer occupancy in a statistical multiplexer with PI traffic. We derived closed form formulas for survivor functions, expected buffer occupancy, and simple asymptotic formula that can be used as a rule of thumb for dimensioning buffer size in designing a statistical multiplexer. The results obtained could shed light on the study of worst case performance of statistical multiplexers for burst-constrained traffic sources in ATM networks.", "doi": "10.7907/pa8m-ay72", "publication_date": "1997", "thesis_type": "phd", "thesis_year": "1997" }, { "id": "thesis:199", "collection": "thesis", "collection_id": "199", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-01162008-075623", "primary_object_url": { "basename": "Minch_ba_1997.pdf", "content": "final", "filesize": 8137114, "license": "other", "mime_type": "application/pdf", "url": "/199/1/Minch_ba_1997.pdf", "version": "v3.0.0" }, "type": "thesis", "title": "Analysis, synthesis, and implementation of networks of multiple-input translinear elements", "author": [ { "family_name": "Minch", "given_name": "Bradley Arthur", "clpid": "Minch-B-A" } ], "thesis_advisor": [ { "family_name": "Mead", "given_name": "Carver", "orcid": "0000-0003-4051-0462", "clpid": "Mead-C-A" } ], "thesis_committee": [ { "family_name": "Mead", "given_name": "Carver", "orcid": "0000-0003-4051-0462", "clpid": "Mead-C-A" }, { "family_name": "Abu-Mostafa", "given_name": "Yaser S.", "clpid": "Abu-Mostafa-Y-S" }, { "family_name": "Barr", "given_name": "Alan H.", "clpid": "Barr-A-H" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Middlebrook", "given_name": "Robert David", "clpid": "Middlebrook-R-D" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "At the time of its invention in the seventeenth century, the logarithmic slide rule literally revolutionized the way calculation was done. From then until the advent of the pocket calculator, this analog computational device was widely used to perform multiplications and divisions, to raise numbers to fixed powers and extract fixed roots of numbers. Today, the slide rule may be gone, but it is not forgotten. In this thesis, I present a class of simple translinear network circuits which essentially function as electronic slide rules, accurately computing products, quotients, powers, and roots. I describe two different analysis procedures that allow us to determine the steady-state relationship between input and output currents. I also describe systematic techniques for synthesizing such circuits whereby we can produce a circuit whose steady-state transfer characteristics embody some desired product-of-power-law relationship between input and output currents. These circuits are made from multiple-input translinear elements; such elements produce output currents that are proportional to the exponential of a weighted sum of their input voltages. We can implement the weighted voltage summations with either resistive or capacitive voltage dividers. We can obtain the required exponential voltage-to-current transformations from either bipolar transistors or subthreshold MOS transistors. The subthreshold floating-gate MOS transistor naturally implements the exponential-of-a-weighted-sum operation in a single device. I will present experimental results from several of these translinear network circuits breadboarded from subthreshold floating-gate MOS transistors. I will also describe and present experimental data from a variety of other implementations of the multiple-input translinear element.", "doi": "10.7907/rh58-rz05", "publication_date": "1997", "thesis_type": "phd", "thesis_year": "1997" }, { "id": "thesis:199", "collection": "thesis", "collection_id": "199", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-01162008-075623", "primary_object_url": { "basename": "Minch_ba_1997.pdf", "content": "final", "filesize": 8137114, "license": "other", "mime_type": "application/pdf", "url": "/199/1/Minch_ba_1997.pdf", "version": "v3.0.0" }, "type": "thesis", "title": "Analysis, synthesis, and implementation of networks of multiple-input translinear elements", "author": [ { "family_name": "Minch", "given_name": "Bradley Arthur", "clpid": "Minch-B-A" } ], "thesis_advisor": [ { "family_name": "Mead", "given_name": "Carver", "orcid": "0000-0003-4051-0462", "clpid": "Mead-C-A" } ], "thesis_committee": [ { "family_name": "Mead", "given_name": "Carver", "orcid": "0000-0003-4051-0462", "clpid": "Mead-C-A" }, { "family_name": "Abu-Mostafa", "given_name": "Yaser S.", "clpid": "Abu-Mostafa-Y-S" }, { "family_name": "Barr", "given_name": "Alan H.", "clpid": "Barr-A-H" }, { "family_name": "Bruck", "given_name": "Jehoshua", "orcid": "0000-0001-8474-0812", "clpid": "Bruck-J" }, { "family_name": "Middlebrook", "given_name": "Robert David", "clpid": "Middlebrook-R-D" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "At the time of its invention in the seventeenth century, the logarithmic slide rule literally revolutionized the way calculation was done. From then until the advent of the pocket calculator, this analog computational device was widely used to perform multiplications and divisions, to raise numbers to fixed powers and extract fixed roots of numbers. Today, the slide rule may be gone, but it is not forgotten. In this thesis, I present a class of simple translinear network circuits which essentially function as electronic slide rules, accurately computing products, quotients, powers, and roots. I describe two different analysis procedures that allow us to determine the steady-state relationship between input and output currents. I also describe systematic techniques for synthesizing such circuits whereby we can produce a circuit whose steady-state transfer characteristics embody some desired product-of-power-law relationship between input and output currents. These circuits are made from multiple-input translinear elements; such elements produce output currents that are proportional to the exponential of a weighted sum of their input voltages. We can implement the weighted voltage summations with either resistive or capacitive voltage dividers. We can obtain the required exponential voltage-to-current transformations from either bipolar transistors or subthreshold MOS transistors. The subthreshold floating-gate MOS transistor naturally implements the exponential-of-a-weighted-sum operation in a single device. I will present experimental results from several of these translinear network circuits breadboarded from subthreshold floating-gate MOS transistors. I will also describe and present experimental data from a variety of other implementations of the multiple-input translinear element.", "doi": "10.7907/rh58-rz05", "publication_date": "1997", "thesis_type": "phd", "thesis_year": "1997" }, { "id": "thesis:5126", "collection": "thesis", "collection_id": "5126", "cite_using_url": "https://resolver.caltech.edu/CaltechETD:etd-12222007-114654", "primary_object_url": { "basename": "Phoong_sm_1996.pdf", "content": "final", "filesize": 10055459, "license": "other", "mime_type": "application/pdf", "url": "/5126/1/Phoong_sm_1996.pdf", "version": "v3.0.0" }, "type": "thesis", "title": "Time-varying and finite field filter banks", "author": [ { "family_name": "Phoong", "given_name": "See-May", "clpid": "Phoong-See-May" } ], "thesis_advisor": [ { "family_name": "Vaidyanathan", "given_name": "P. P.", "clpid": "Vaidyanathan-P-P" } ], "thesis_committee": [ { "family_name": "Vaidyanathan", "given_name": "P. P.", "clpid": "Vaidyanathan-P-P" }, { "family_name": "Bruck", "given_name": "Jehoshua", "clpid": "Bruck-J" }, { "family_name": "Effros", "given_name": "Michelle", "clpid": "Effros-M" }, { "family_name": "McEliece", "given_name": "Robert J.", "clpid": "McEliece-R-J" }, { "family_name": "Xia", "given_name": "Xiang-Gen", "clpid": "Xia-Xiang-Gen" } ], "local_group": [ { "literal": "div_eng" } ], "abstract": "Filter banks find many applications in signal processing. This thesis deals with four different problems in filter banks.\r\n\r\nFirst we find a new application of filter banks: Filter bank convolver. We prove two filter bank convolution theorems which tell us how to do the convolution in the subbands. Applying the multirate technique to the problem of convolution, we obtain a significant improvement in the accuracy of the convolutional result when the computation is done with finite precision. The derivation also leads to a low sensitivity robust structure for FIR filters.\r\n\r\nIn the second part, a new class of two-channel biorthogonal filter banks is proposed. We successfully design IIR filter banks which achieve the following desired properties simultaneously: (i) Perfect reconstruction (PR); (ii) causality and stability; (iii) near linear-phase; (iv) frequency selectivity. Two classes of causal stable maximally flat IIR wavelets are derived and closed form formulas are given. We also provide a novel mapping of the proposed 1D framework into 2D. The mapping preserves: (i) PR; (ii) stability in the IIR case and linear phase in the FIR case; (iii) frequency selectivity; (iv) low complexity.\r\n\r\nIn the third part, the theory of paraunitary (PU) filter banks is extended to the case of GF(q) with prime q. We show that finite field PU filter banks are very different from real or complex PU filter banks. Despite all the differences, we are able to prove a number of factorization theorems. All unitary matrices in GF(q) are factorizable in terms of Householder-like matrices. The class of first-order PU matrices, the lapped orthogonal transform in finite fields, can always be expressed as a product of degree-one or degree-two building blocks.\r\n\r\nFinally the theory of conventional LTI filter banks is extended to the time-varying case. We develop a polyphase representation method for time-varying filter bank (TVFB). Using the proposed polyphase approach, we are able to show some unusual properties which are not exhibited by the conventional LTI filter banks. For example, we can show that for a PR TVFB, the losslessness of analysis bank does not always imply that of the synthesis bank, and a PR TVFB in general will only generate a discrete-time frame, rather than a basis, for the class of finite energy signals. The class of lossless TVFB is studied in detail. We show that all lossless linear time-varying systems are invertible and provide explicit construction of the inverse. The interplay between invertibility, uniqueness and losslessness of the inverse is investigated. The factorizability of lossless TVFB is addressed and we show that there are factorizable and unfactorizable examples.", "doi": "10.7907/1bzz-ba31", "publication_date": "1996", "thesis_type": "phd", "thesis_year": "1996" } ]