[ { "title": "An Electrodynamic Perspective of Black Holes", "type": "thesis", "publication_date": "2025", "doi": "10.7907/ypb6-3e47", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:05292025-212148506", "abstract": "Numerical relativity (NR) is a powerful modeling tool for the dynamics of general-relativistic systems that are difficult to analyze analytically. Prior work has led to a tetrad formulation of the 3+1 decomposition of the Einstein field equations (EFE) that bears a striking resemblance to electrodynamics, done by recasting Einstein\u2019s equations into a set of coupled nonlinear Maxwell equations. We use gravitational electric and magnetic fields developed from this theory to analytically probe Schwarzschild and Kerr solutions to the EFE. We compare the resulting Kerr dynamics with a numerical simulation of the Kerr spacetime. We then extend this analysis to visualize the inspiral, merger, and ring-down of a binary black hole collision simulated in moving puncture gauge.", "author_list": "Boyeneni, Siddharth" }, { "title": "Improving Parameters of Asymptotically Good Quantum LDPC Codes via Stronger Product Expansion", "type": "thesis", "publication_date": "2025", "doi": "10.7907/665j-dr37", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:06062025-083109104", "abstract": "Quantum low-density parity-check (qLDPC) codes are a promising path toward scalable, fault-tolerant quantum computation. This thesis focuses on improving the relative distance of asymptotically good qLDPC codes, with a particular emphasis on quantum Tanner codes. We present a refined analysis of product expansion in tensor codes and introduce a stronger form of the expansion property that leads to improved lower bounds on code distance. Numerical results further illustrate how our method enables improved trade-offs between code parameters under practical constraints. While our analysis is framed in the quantum Tanner code setting, the techniques are broadly applicable to other constructions whose local codes are based on tensor product decompositions. Our work contributes to closing the gap between asymptotic constructions and realizable quantum codes.", "author_list": "Cai, Yiyi" }, { "title": "Physically-Motivated Modeling of Kinetic Inductance Phonon-Mediated Detector for Light Dark Matter Searches", "type": "thesis", "publication_date": "2025", "doi": "10.7907/nd6z-1y53", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:06172025-193756039", "abstract": "The properties of dark matter (DM) is one of the most exciting mysteries in astrophysics, and they are important in understanding cosmological structure formation and could potentially reveal new physics. Direct searches for DM necessitate using ultra-sensitive quantum sensors, one of which is the kinetic inductance phonon-meditated detector (KIPM). Understanding KIPM response is vital to understanding the device's energy resolution. Here, we present a physically-motivated model of KIPM response based on quasiparticle and phonon lifetimes. We examine its adherence to experimental data in three formulations, which either six, five, or four time constants. We examined the temperature-dependence of these time constants and compare them to the results of previous models. All three models fit to data at below 75 mK, with successful fits up to 150 mK in some cases; the five time constants model presents the closest match of temperature-dependence of quasiparticle and phonon lifetimes to existing knowledge, while goodness of fit indicates that the six time constant model have the potential to fit high temperature data better. This paper detailed both the behaviors of the physically-motivated models as well as fitting considerations for the behaviors of the fit.", "author_list": "Cap, Chi Lan" }, { "title": "Methodology and Insights for System Calibration in Multi-Angle Illumination Imaging", "type": "thesis", "publication_date": "2025", "doi": "10.7907/g6km-ac77", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:08132025-220508308", "abstract": "Multi-angle illumination-based computational microscopes have emerged as a promising class of imaging systems due to their capabilities and robustness across a wide range of applications, from biological imaging to materials inspection. In particular, quantitative phase imaging methods such as Fourier Ptychography Microscopy, Angular Ptychographic Imaging with Closed-form solutions and Kramers-Kronig relations leverage multi-angle illumination to surpass traditional space-bandwidth limitations and digitally correct aberrations. However, the performance of these systems is highly sensitive to misalignment in the illumination angles, and even minor perturbations can significantly degrade reconstruction quality and necessitate time-consuming recalibration. Thus, there is a pressing need for efficient and robust illumination angle calibration in such imaging modalities. We investigate how angular misalignments affect reconstruction fidelity and systematically evaluate a range of digital calibration strategies, including classical geometric models, cross-correlation-based methods, and learning-based approaches. These methods are benchmarked across varying signal levels and sample types. Our findings offer practical insights into selecting and deploying robust calibration techniques, ultimately supporting more resilient, reproducible, and high-throughput computational microscopy systems.", "author_list": "Deng, Catherine" }, { "title": "The Assembly and Testing of the Spin Dressing Magnet for the Neutron Electric Dipole Experiment", "type": "thesis", "publication_date": "2025", "doi": "10.7907/rjne-gy68", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:06112025-032822492", "abstract": "The discrepancy between the quantity of matter and anitmatter in the universe is something that can likely be attributed to violations in the fundamental symmetries of the universe; however, much like the antimatter itself, there is a discrepancy between the required versus obeserved magnitude of these violations. One theory states that, to account for these violations in symmetry, the neutron must have an electric dipole moment. One such method to find the existence and magnitude of the neutron electric dipole moment (nEDM) is the critical dressing method. Such a method requires the use of two superconducting magnets with perpendicular magnetic fields. This specific method of critical dressing uses superfluid Helium-4, polarized Helium-3, and ultracold polarized neutrons, with critical dressing occurring when the Helium-3 precession rates are equivalent. This method is used to determine the existence of an nEDM, if there is critical dressing with an electric field, there is no nEDM, but if there is a precession rate difference with the electric field, there is an EDM that can thus be measured. Over the past several months, the assembly of the spin dressing magnet used in the critical dressing portion of the nEDM experiment has begun. This has included assembling the boss rings, constructing the magnet frame, placing the story sticks and wire guides, and winding the superconducting wire around the coil skeleton. Data was also taken using this wire. Furthermore, simulations have been run on COMSOL Multiphysics to compare the theoretical predictions with the measurements of the magnetic field and B-field gradients produced by the spin dressing magnet.", "author_list": "Fox, Jessica Lauren" }, { "title": "Investigating the Biological Mechanism of N\u2082O Emissions from Arid Southern Californian Drylands", "type": "thesis", "publication_date": "2025", "doi": "10.7907/9a4y-mm41", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:06122025-192128192", "abstract": "Nitrous oxide (N\u2082O) is a powerful greenhouse gas, each molecule capable of warming the atmosphere 273 times more effectively than CO\u2082. Arid soils that have been rewetted by rainfall events can produce some of the highest instantaneous N\u2082O emission rates recorded globally. Recent work has shown that the majority of these emissions are biologically produced. While these emissions have classically been attributed to bacterial and fungal denitrification catalyzed by catabolic nitric oxide (NO) reductases (e.g. NOR), measured N\u2082O isotopic fingerprinting (site preference, SP) more closely matches flavohemoglobin enzymes involved in nitric oxide detoxification (e.g. Fhp). Analysis of the microbial community of the site demonstrates that fhp is significantly more phylogenetically abundant than nor. We hypothesize that NO detoxification pathways are responsible for the initial pulse of N\u2082O production after rainfall, with denitrification only becoming dominant after a few hours. N\u2082O production is only triggered once some critical saturation with the water is reached, suggesting that the soil community has to receive enough water to become anaerobic. Using coupled measurements of oxygen and N\u2082O concentration in soils, we show that N\u2082O production begins only once the added water depletes the soil of oxygen. Initial measurements of N\u2082O production from Pseudomonas synxantha, a bacterium isolated from soil, demonstrate clear differences in the timing and quantity of gas production following rewetting via the detoxification and denitrification pathways. We thus suggest that previously overlooked detoxification pathways may play key roles in observed biogeochemical events, as appears to be the case with soil N\u2082O emissions.", "author_list": "Isella, Emma Xueqian" }, { "title": "Batik: a Vision Language Model for End-to-End Social Behavior Discovery, Interpretation and Annotation", "type": "thesis", "publication_date": "2025", "doi": "10.7907/cmvx-1n97", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:06062025-053817901", "abstract": "Quantitative analysis of animal behavior is a burgeoning field. By converting behavior into measurable features, the field replaces anecdotal observations with precise, data-driven insights into how animals interact with their environment and with one another. Through certain analyses, reproducible structure and diversity within behaviors are revealed, illuminating complex behavioral patterns. Most current state-of-the-art methods focus on annotation and segmentation of behavior using pose-estimation; these methods attach nodes to body parts of mice which then compute a features space. This feature space is then used for discovery of behavior classes or training supervised behavior classifiers. However, this excludes the time-consuming task of interpreting resulting behavioral syllables and has multiple failure modes, such as an inability to attend to frames where there are other objects of interest or frames where the nodes are all on top of one another. The majority of these methods use a convolutional neural network structure. In recent years, a new set of feed-forward neural networks called transformers have been proven to surpass CNNs on most vision-related tasks. Batik addresses the long time-commitments of interpreting these syllables by using multimodal transformers to extract unsupervised features directly from raw video, and perform end-to-end analysis, bypassing pose estimation. Alongside state-of-the-art supervised annotation, Batik leverages fine-tuned large language models to automate discovery and provide expert human-level interpretation of behavior syllables, offering researchers a transformative UI-based tool for behavioral analysis through vision-language models. Through these methods, we show a 96% accuracy for syllables like attack and mount, a large jump from previous methods (85%). We also accurately identify differences in behavior in different metabolic states, as well as an interpretation with sub behaviors for the broad investigative behavior. We further apply our method to other species datasets, correctly classifying distinct fly aggressive behaviors with no additional fine-tuning of the underlying model, showing our model\u2019s generalizability.", "author_list": "Kolhe, Rohan Rajendra" }, { "title": "Constraining the Distribution of 3D Fractal Structures in Mud Flocs", "type": "thesis", "publication_date": "2025", "doi": "10.7907/cdeh-e474", "cite_using_url": "https://resolver.caltech.edu/CaltechTHESIS:06112025-081128769", "abstract": "Mud builds coastal landscapes and governs the long-term evolution of river deltas, floodplains, and estuaries, yet predicting its transport remains difficult because mud aggregates into flocs with complex, fractal structures that deviate from simple particle behavior. The three-dimensional (3D) fractal dimension of these flocs sets their settling and sediment transport characteristics, but reliably determining this parameter across diverse environments is a persistent challenge. Conventional aggregation of floc data often obscures real structural diversity and can yield misleading fractal dimensions due to Simpson\u2019s Paradox. This study tests the hypothesis that stratifying settling data by image-derived two-dimensional (2D) fractal dimension enables more accurate inference of the hydrodynamically relevant 3D fractal dimension. Controlled experiments with freshwater flocs, formed under varied shear and particulate organic matter (POM) conditions, were conducted using in-situ imaging, PIV-corrected tracking, and box-counting analysis to resolve structural differences. Results demonstrate that aggregation overestimates the 3D fractal dimension, while stratification reveals clear trends: the inferred 3D fractal dimension increases with shear stress and decreases with particulate organic matter content. These findings provide a basis for more realistic floc modeling and improve predictions of fine sediment transport.", "author_list": "Noh, Sangwon (Brayden)" } ]