80264 research outputs found
Sort by
Ultrastrong Vacuum–Matter Interactions in Semiconductors and Magnets
No full textThere is growing interest in using optical cavities to uncover new phases and phenomena, relying solely on vacuum electromagnetic fields enhanced within the cavity, without any external fields. Ultrastrong coupling (USC) between vacuum and matter is a prerequisite. USC occurs when the vacuum–matter coupling rate becomes a significant fraction of the bare frequencies of the systems. In this dissertation work, we studied the USC of vacuum and matter in semiconductors and magnets. First, we realized multimode phonon-polaritons in lead halide perovskites. The USC of two optical phonon modes with the vacuum induced novel vibrational properties. We also demonstrated via photoluminescence measurements that electron–phonon interactions are modified in this system. Next, in a quantum Hall system, we demonstrated the breakdown of the electric-dipole approximation using nanoscale cavities, revealing forbidden electronic transitions. We further investigated quantum–classical correspondence in the nonlinear regime, which was explained by a quantum model. Finally, we observed a magnonic superradiant phase transition in an ErFeO3 crystal. USC between an Fe3+ magnon mode and an Er3+ electron paramagnetic resonance resembled vacuum–matter interactions, leading to the phase transition. These findings provide quantum optical strategies for creating and controlling novel phases in condensed matter via control of the quantum vacuum surrounding the matter
1.4 Nature Conservation in the Age of Bioengineering
Full text linkHighlights how bioengineering could positively contribute to conservation efforts through thoughtful, collaborative, and ethical approaches.This entreaty was created as part of The Spirit of Asilomar and the Future of Biotechnology summit (February 23-26, 2025) in Pacific Grove, CA.This entreaty explores the promising role of bioengineering in constructively supporting conservation efforts through responsible innovation and interdisciplinary collaboration. It outlines eight concise statement points that reflect shared principles and opportunities for action
Flash Joule Heating for Materials Production
No full textFlash Joule heating has been widely used as an ultrafast, scalable, and versatile synthesis
method, most prominently in the synthesis of flash graphene and other carbon materials.
Whereas most chemical synthesis methods transfer heat through a medium into which most heat
is lost, flash Joule heating reactions utilize the target feedstock itself as the heating medium,
enabling near optimal heating efficiency and consequently extremely high heating rates. Herein,
I present an overview of the use of flash Joule heating for materials production, including
graphene, graphite, carbon nanotubes, doped graphene, silicon carbide, and p-block metal
dichalcogenides. I present different engineering and reaction techniques to facilitate the
kilogram-scale production of these materials while performing life cycle assessments and
technoeconomic analyses of these processes. I further highlight the impact that the passage of
electrical current through the reactant feedstock has on the mechanics of the flash Joule heating
technique, finding that this phenomenon can reduce reaction activation energy. I finally discuss
the historical foundations of graphene production and provide evidence that Thomas Edison may
have synthesized graphene as early as 1879
Essays in Industrial Organization: Cartels and Competition
No full textIn Chapter 1, I examine the impact of a market allocation cartel in the Japanese electricity retail market, active from 2018 to 2020. Four incumbents restricted competition during this period by avoiding entry into each other’s regions.
Analyzing electricity procurement auctions from both competitive and cartel periods, I find that cartel members reduced their participation rates and submitted complementary bids in other regions while increasing bid levels within their own regions, leading to winning bids rising by up to 7.7\% despite continued competition with non-cartel firms.
Counterfactual simulations using a model of auctions with asymmetric, risk-averse bidders suggest that without the cartel, continued market entry by these firms could have lowered winning bids by up to 5.5\% and reduced winning costs by 3.4\%.
While increased competition led to minor inefficiencies due to asymmetry among bidders, the cartel’s exclusionary practices caused inefficiencies of up to 26\%.
Additionally, shifts toward nuclear energy generation were associated with lower procurement costs.
In Chapter 2 (with Suguru Otani), we revisit conduct parameter estimation in homogeneous goods markets to resolve the conflict between Bresnahan (1982) and Perloff and Shen (2012) regarding the identification and the estimation of conduct parameters.
We point out that Perloff and Shen’s (2012) proof is incorrect, and its simulation setting is invalid.
Our simulation shows that estimation becomes accurate when demand shifters are properly added in supply estimation and sample sizes are increased, supporting Bresnahan (1982).
In Chapter 3 (with Suguru Otani), we revisit the identification result of conduct parameters in homogeneous goods markets by Lau (1982).
Lau's result is that the conduct parameter cannot be identified if and only if the demand function is separable but not a specific functional form.
We show that the result is incorrect by providing a separable demand function that induces identification.
This implies that the class of inverse demand functions that lead to identification is broader than Lau (1982) considers.
Therefore, the conduct parameter can be identified in the broader class of inverse demand functions when the market has enough variation
Perturbation-driven responses across scales in Hydra: How a simple system handles change
No full textBiological systems face the challenge of remaining functional amid internal and external change. Understanding how organisms respond to perturbations enables us to dissect how structure, state, and function are related. Hydra vulgaris, with its simple anatomy, continuously remodeling body, and diffuse nervous system, offers a tractable model for answering these questions. In this thesis, we examine Hydra’s responses to environmental and internal state changes, as well as intrinsic differences in stability across cell types. We first show that Hydra’s neural responses to thermal stimulation are robust, remaining functionally intact despite two-fold changes in the number of neurons. We then demonstrate that Hydra adapts to internal states such as satiety, which alters the phototactic behavior. We determine that differences in locomotion rate are sufficient to explain this satiety-driven phototaxis. Finally, to understand how internal state shapes the molecular landscape, we develop the first single cell multiomic atlas in Hydra (joint scRNA-seq and scATAC-seq) to provide layered insights into gene regulation. By examining perturbation-driven effects across biological scales in Hydra vulgaris, this work provides a framework for understanding how simple biological systems remain functional under external and internal disruptions
Analysis for Science Librarians of the 2024 Nobel Prize in Physics: Foundational Discoveries Enabling Machine Learning with Artificial Neural Networks
Full text linkThe Royal Swedish Academy of Science has awarded the Nobel Prize in Physics for 2024 to two scientists, John Hopfield and Geoffrey Hinton, for their foundational work developing models for associative memory and artificial neural networks, respectively. This article provides an analysis of the citation and publication patterns of each awardee after a brief introduction to the science and scientists who received the prize
Personalizing Assessment of Motor Impairment for Stroke Rehabilitation
No full textMotor impairment assessments of stroke are used by therapists to track recovery and prescribe treatment protocols that optimize rehabilitative outcomes. For outpatient-based stroke rehabilitation, lengthy administration times of traditional clinical assessments limit associated benefits and preclude additional therapist-directed rehabilitation that improves outcomes. Robotic and sensor-instrumented systems provide an objective method of assessment that offers additional resolution via measurement of kinematic quantities of movement. Prior research has validated assessment automation with such systems, but has neglected automating assessment using systems that are independently usable by stroke survivors. In this thesis, we analyze the effect of range of motion on a common assessment metric, movement smoothness, which gauges motor coordination as a facet of motor impairment. Based on our findings, we present guidelines for implementation of movement smoothness assessment that preserves task construct validity. A novel device, the FlexWrist, is presented as a usability-focused, glove-based flex sensor system for recording home exercise movement of stroke survivors' hemiparetic wrist and hand
Influence of Auditory Information on Drivers’ Abilities to Extrapolate Motion of a Lead Vehicle in a Traffic Environment
Full text linkNearly 29 percent of all traffic collisions between two vehicles on the road are rear-end collisions, primarily caused by shortened headway. This shortened headway could be attributed to a driver’s inability to accurately extrapolate motion. Components of music, such as tempo, may affect drivers' abilities to extrapolate motion, but it has not yet been studied. This is important to understand, as about 70% of Americans listen to music while driving. Therefore, the current research aimed to understand whether music, specifically the tempo, affects drivers' abilities to extrapolate motion. Drivers watched simulated car-following scenes and either heard no sound or a sound with a slow or fast tempo. After an interruption, the scene reappeared with the lead vehicle at the correct position or at a more or less advanced position. Drivers reported whether the lead vehicle reappeared at the correct position. The first experiment examined if drivers' abilities to extrapolate motion during a 777 ms interruption varied based on the type of sound played while the driver drove the same speed, faster, or slower than the lead vehicle (i.e., relative velocity). The second experiment examined if the timing of sound (played during the entire scene including the interruption, only before the interruption, or only during the interruption) influenced motion extrapolation and whether a 3.2 s interruption duration affected drivers' reappearance judgments when the driver drove faster than the lead vehicle. Motion extrapolation at an interruption of 777 ms was based only on the relative velocity of the vehicles in the scene. The effect of tempo was not significant. When the interruption was 3.2 s, the tempo influenced drivers' abilities to extrapolate motion. Specifically, judgments were more biased towards advanced reappearance positions at a faster tempo compared to a slower tempo. When the interruption duration was 3.2 s but not 777 ms, drivers relied on the auditory stimuli, resulting in a greater difference in reappearance judgments among sound conditions. Thus, auditory information can affect drivers' abilities to extrapolate motion, but only at relatively longer interruptions
Fine-Grained Paging Mechanism for Offloading-Reloading Tensor for LLM
No full textThe rapid growth in the size and complexity of large language models has imposed
severe challenges on memory management, particularly when these models are
deployed on GPUs with limited memory. This thesis introduces a fine-grained paging
mechanism that dynamically offloads and reloads tensors at the granularity of
individual operations, thereby mitigating out-of-memory (OOM) issues during the
inference and prefill phase of transformer-based models. Instead of traditional static,
layer-based offloading methods, the proposed approach uses compile-time, simulationbased
memory allocation to optimize GPU memory usage, making runtime possible
under severe memory constraints.
This work is based off of the Einsummable system, a framework that represents
tensor computations using Einstein summation notation. Einsummable transforms
high-level mathematical specifications into an optimized execution pipeline through a
series of intermediate representations, notably the taskgraph and the memgraph.
The taskgraph captures the data dependencies and operational flow of tensor computations,
while the memgraph extends this representation by incorporating detailed
memory location information and managing offload-reload operations. The transformation
from taskgraph to memgraph is achieved through a simulated execution
process — the core of this thesis — that relies on two key components: an allocation
horizon, which pre-allocates memory for future operations, and an execution horizon,
which tracks the simulated execution progress of the computation.
A key contribution of this thesis is the design and implementation of specialized
memory allocation routines—simMalloc, simMallocForceReld, and simMallocOffld.
These routines not only allocate memory for tensor outputs but also manage dependencies
by inserting offload and reload nodes into the memgraph whenever GPU
memory resources depletes. By leveraging full knowledge of the simulated execution
order, our offload-reload heuristic selects tensors for offloading based on their
computed reuse distance, thereby deferring memory transfers until they are most
convenient. This future-aware strategy mitigates the frequency and impact of memory
transfers compared to reactive approaches, enabling a finer control over GPU
memory usage.
Extensive experimental evaluations were conducted using two configurations of
NVIDIA GPUs — Tesla P100 and V100 — to benchmark the performance of the
proposed system against state-of-the-art techniques such as ZeRO-Inference. The
evaluation focused on the prefill stage of inference in LLaMA models with 7B and 65B
parameters, a phase known to be particularly memory-bound. The results demonstrate
that the fine-grained paging mechanism supports a broader range of configurations,
successfully executing inference tasks across varying batch sizes and sequence
lengths. While the finer granularity of tensor-level management introduces some
communication overhead due to more frequent offloading and reloading, the overall
improvements in memory utilization and reduction in OOM errors outweigh these
costs.
In summary, this thesis makes a contribution to the field of deep learning by
addressing the critical challenge of GPU memory constraints through a fine-grained
paging mechanism. Future work will explore further optimizations to reduce communication
overhead, overall computation latency, and GPU RAM utilization
Cheap Meat, Plastic Lives: Cattle Imports, Veterinary Care and Shifting Livestock Economies in Turkey
No full textThis dissertation explores Turkey’s "cheap meat policy" and the traffic in cattle through an ethnographic study of live animal imports. It investigates how cattle are turned into globally traded commodities by focusing on two interconnected dimensions: the sensorial frictions of global live trade and the longer histories of (forced) animal mobilities. The research centers on underexplored sites of the Animal Industrial Complex (Noske 1989), such as importing companies, ports, customs areas, and agricultural expos, where cattle are (re)made into globally tradable commodities. It emphasizes the importance of sensory experience, especially touch and smell, in mediating and at times disrupting the smooth functioning of the traffic in cattle, which situates live imports not just as an economic exchange, but a system that reorganizes and consolidates certain roles ascribed to cattle, veterinarians, importers, and state representatives in their respective ways. By foregrounding the labor, infrastructure, and political strategies required to sustain import-driven livestock economies, the dissertation highlights tensions between economic imperatives and the embodied, sensory presence of living beings in global trade. Rather than approaching “cheap meat” through consumption or as a given category, it focuses on the political-economic processes that situate it as a reflection of claims over welfare and development, which defer conversations around agricultural transformations in Turkey, international regulatory frameworks, and internal conflict and displacement. These longer histories shape the organization of live trade and reveal continuities in global systems of forced mobilities and exploitation, exemplified by live carrier ships functioning as “floating barns.” While these recurrent histories underscore the exhaustive powers of animal capitalism, they also introduce frictions, failures, and resistances that challenge capital’s totalizing grip