Washington University Medical Center
Washington University St. Louis: Open ScholarshipNot a member yet
26344 research outputs found
Sort by
New Signal Identification Algorithms for Enhanced Gamma-Ray Burst Detection in the Advanced Particle-Astrophysics Telescope
Get PDFThis work presents a series of algorithmic advancements aimed at improving photon signal identification and gamma-ray burst (GRB) source localization for the Advanced Particle-astrophysics Telescope (APT) and its Antarctic Demonstrator (ADAPT). These advancements are aimed at identifying valid signals in noisy environments. Previous methods failed to effectively distinguish real photon signals from noise, prompting us to develop a new photon detection algorithm with peak counting. Instead of integrating all waveform data in the observation window, we use multiple thresholds to accurately identify single-photon and two-photon arrival events, minimizing false counts due to amplifier noise. The new peak count algorithm also incorporates extra logic to avoid repeated counting and improve accuracy. Additionally, we address challenges in optimizing the integration window to mitigate dark count effects. Our strategy is informed by analysis based on the distinct statistical distributions of real photons and dark counts. We found the minimum standard deviation integration window for different numbers of photon arrival events. In the digitizer stage, a new filtering mechanism ensures events possess enough active detector layers to enhance Compton reconstruction performance. Raw sensor signals are formed into ``hit groups\u27\u27 that represent individual gamma-ray interactions in the scintillator. Previous algorithms depend upon multiple thresholds for forming hit groups. To further improve hit group detection, especially under high zero threshold 2 constraints, we introduce the concept of an ``extra hit group. We only implement this extra hit group when we can not identify a normal hit group that only uses adjacent data. Combining non-adjacent data into an extra hit group when no normal hit group is identified reduces the loss of real photon events. These combined enhancements contribute to more robust and accurate localization of GRB sources, especially when dealing with background noise
Transfer Learning for Temporal Logic Objectives
Get PDFReinforcement learning algorithms can enable autonomous systems to learn the control skills needed to accomplish a task specified by a linear temporal logic formula. However, they cannot be transferred to a new task, even when the two are very similar. For each new task, the policy must be redesigned from scratch, which is a common limitation of existing reinforcement learning methods for temporal logic tasks. A proposed solution to this problem leverages the similarity between past and new tasks to reuse already learned control skills to accomplish the new task, with minimal or no retraining.
Rather than learning a single policy that satisfies a task specified by a linear temporal logic formula, this approach decomposes the formula into reach-avoid subtasks, where policies are instead learned for each reach-avoid subtask and stored in a library of reusable control skills. Similarity metrics between two formulas are defined and algorithms for their computation are proposed. These are essential for determining if the library of reusable control skills contains the necessary policies to accomplish the new task. Finally, a simulation of an autonomous system that was trained to satisfy a temporal logic task is presented completing a new, unseen temporal logic task without any retraining
Development of a Light-Based Platform to Assess Cervical Mucus Barrier Function and Physical Mucus Properties
Get PDFPreterm birth is the leading cause of neonatal morbidity and mortality worldwide, yet current clinical tools for predicting its risk lack specificity and reliability. Cervical mucus plays a crucial role in maintaining pregnancy by forming a barrier that protects the uterus against infection. This study investigates the use of laser speckle rheology as a noninvasive technique to assess cervical mucus biomechanics. Synthetic cervical mucus was fabricated at varying hydration levels and tested using a conventional rheometer that confirmed that the mechanical properties of the synthetic mucus samples were representative of physiological cervical mucus. These results were compared to the complex moduli found using the laser speckle rheology system, finding promising agreement between the two methods. These findings establish laser speckle rheology as a potential noninvasive method for characterizing cervical mucus mechanical properties. By providing quantitative insights into cervical mucus integrity, LSR holds promise as a diagnostic tool for PTB risk assessment. Future work will focus on refining optical property calibration, optimizing clinical implementation, and validating LSR against a broader cohort of biological samples. Ultimately, this technology aims to enhance early PTB detection strategies, leading to improved pregnancy outcomes
Direct Computations of Spatially Resolved Viscoelastic Moduli of Biomolecular Condensates
Get PDFBiomolecular condensates are viscoelastic materials formed by liquid-liquid phase separations (LLPS) of biopolymers. In this study, we develop a modified graph Laplacian-based collective model to characterize viscoelastic heterogeneity within condensates based on results of lattice-based Metropolis Monte Carlo (MMC) simulations. By integrating random graph models and simulations of A1-LCD, a type of intrinsically disordered protein, we examine how network topology influences storage modulus, loss modulus, crossover frequency, and relaxation time spectra. Our results reveal a strong correlation between topological features and mechanical response, with the condensate interior exhibiting higher stiffness and faster relaxation than the interface. The relaxation spectra provide rich insights into dynamic behavior beyond what moduli alone can capture. This framework offers a scalable, interpretable approach for quantifying spatial viscoelasticity in biomolecular assemblies
Functional Devices Based on Freestanding 2D Materials
Get PDFTwo-dimensional (2D) materials have attracted extensive attention in the field of nanoelectronics due to their atomic-scale thickness, high surface-to-volume ratio, tunable electronic properties, and compatibility with low-temperature processing. These characteristics make them highly suitable for the construction of emerging device architectures, particularly in both ionic and electronic devices.
In this work, we investigate the application of 2D materials in two distinct classes of devices: ionically-driven memristors and electronically-dominated metal–semiconductor contacts. For the memristor study, we fabricate heterostructure-based resistive switching devices using h-BN and WSe2 as active layers. These 2D material-based memristors exhibit stable power consumption loops and high linearity under pulse modulation, demonstrating their potential as synaptic elements for neuromorphic computing and energy-efficient memory systems.
In parallel, we explore the role of 2D materials in contact engineering by analyzing the interface between MoS2 and various metal electrodes. Systematic experiments reveal that the selection of contact metals and tuning of work function critically influence the Schottky barrier height and carrier injection efficiency, offering valuable insights into the optimization of 2D semiconductor-based transistors.
Collectively, this study demonstrates the versatility and promise of 2D materials for both ionic and electronic device applications, highlighting their potential for integration into next-generation computing architectures
Froggy People: Anthropomorphism and Slice-of-Life in Autobiographical Comics
Get PDFThis essay explores how the following three comics categories overlap and interact with one another: slice-of-life, anthropomorphic, and autobiographical. I search for reasons why comic makers would choose to tell their own daily life stories using animal avatars rather than their own faces. While each of these three categories has been studied extensively individually, I hope to find additional insights by observing how comic makers employ the genres together as storytelling devices. Each category has unique narratological power which impacts the reader’s experience. For example, anthropomorphism affects how the author and reader internalize and relate to characters. Slice-of-life comics are indicative of cultural performance. And, autobiographical comics communicate a sense of trust between the reader and author about what is true. I examine these attributes by analyzing the work of comic artists who play in between these genres, with the help of frameworks laid by comic and visual culture scholars. Ultimately, I demonstrate how the combination of these genres can have useful narrative potential
Electrochemical Performance Analysis of a URFC in Fuel Cell Operation Under Variable Flow and Backpressure Conditions
No full textUnitized regenerative fuel cells (URFCs) offer a compelling energy storage solution by integrating both fuel cell and electrolyzer functionalities into a single device. This study evaluates the electrochemical performance of a URFC system running in fuel cell mode utilizing bifunctional Pt-RTO and Pt/C catalysts and Nafion membranes, tested under varying flowrates and backpressures. Linear sweep voltammetry (LSV) measurements were performed using a Scribner URFC test station to extract maximum current density and characterize operational trends. Across the full range of experiments, increasing oxygen flowrate and applying moderate to high backpressure improved maximum current densities, while low or zero backpressure significantly reduced performance. The highest current density observed was 119.1 mA/cm2 under 10 psi backpressure and 100 mL/min oxygen flowrate. These findings provide insight into mass transport and catalyst utilization challenges in URFCs and inform future strategies for optimizing device performance and reliability. Recommendations for future work include mathematical modeling, durability testing, and mechanical improvements to membrane-electrode assembly fabrication
The Illusion of Inclusion: The False Promise of the New Governance Project for Content Moderation
Get PDFBecause private companies now control the most prominent communication platforms, the most pressing question in the field of content moderation is how to ensure that the governance of public discourse responds to public values. The prevailing approach, given that the state cannot regulate speech directly, is that state regulation can be substituted with audited self-regulation, broad stakeholder participation, and negotiated rulemaking. In this model, which this article refers to as the “new governance model for content moderation,” companies include advocates as representatives of the public in their processes to govern online speech. Ideally, they negotiate policy goals and share responsibility for achieving them. The end goal is to have a process in which public values are given effect. This article argues, however, that this governance model is unsound in both theory and practice. In the field of content moderation, the ambition of constructing public values through a collaborative process between companies and stakeholders is conceptually incoherent: those interests that cannot elicit the cooperation from corporate actors and are not consistent with the values of participating advocates are excluded by design. In practice, no present or past demonstrations have shown that the inclusion of advocates in speech governance and the agreements they reach with companies have epistemic credibility to construct the public interest. Though those flaws might seem unsurprising, scholars and activists double down on independence, diversity, and expertise as design strategies that can result in self-regulatory bodies that could adequately set policy goals. This article advocates for pluralism as a framework that more effectively achieves the participatory goals of new governance. It argues that the state has a central role to play in creating a plural and contested public sphere. A robust legal system can complement self-regulation and push it structurally in the direction of public values
Computational Complexity of Soundness Verification for Neural Networks
Get PDFNeural networks are an increasingly ubiquitous tool in systems of varying complexity across a range of domains. While these tools can be used to learn and predict complex functions, their opaque nature limits the scope of their acceptable applications. In particular, a lack of performance guarantees means that they are unsuitable for safety-critical applications such as self-driving cars and scheduling systems. Neural networks trained to solve NP-complete problems, in particular, are unlikely to be able to solve the problem exactly. However, a weaker soundness guarantee may be sufficient for some systems, e.g., that positive instances of the problem may be rejected but no negative instances are accepted. Research into the complexity of securing such guarantees would improve our understanding of the applications to which neural networks can be effectively and economically applied.
In this work, we show that the verification of soundness for neural networks trained to solve 3SAT formulae is a ΠP2-complete problem and conjecture that the same hardness holds for neural networks solving other NP-complete problems. Therefore, it is unlikely that even these weaker guarantees on the performance of neural networks solving hard problems can be economically secured. A related result drastically expands the set of problems known to be complete for higher levels of the polynomial hierarchy. These problems are formulaically constructed in terms of problems complete for lower levels of the polynomial hierarchy
Landscapes of Translation: Centering Nâzım Hikmet’s Translators in World Literature
No full text*Landscapes of Translation* begins with a simple question: Who were the English and French translators of Turkey’s “world poet” Nâzım Hikmet? I construct a translation history for Nâzım by humanizing the lives of his translators, starting with his earliest mediators, like Nermin Menemencioğlu and Fikret Adil, through his Cold War champions like Sabahattin Eyüboğlu, Nilüfer Mizanoğlu Reddy, and Rosette Coryell. I then turn to those who brought his epic-novel-in-verse, *Memeleketimden İnsan Manzaraları* (Human Landscapes from My Country) to the world: Münevver Andaç, Taner Baybars, Mutlu Konuk, and Randy Blasing. By historicizing their motivations, ambitions, and networks, I demonstrate how his translators were the chief architects of his international stature, centering them in the formation of world literature to model a biographical approach for the global circulation of poetry. I then consider Nâzım’s translators in the broader context of Turkish poetry circulation by reconstructing the history of the first anthology of modern Turkish poetry in English, Derek Patmore’s *The Star and the Crescent*, describing its implicit sponsorship by the British Council and its legacy of defining Turkish poetry in translation through representative anthologies that often excluded Nâzım. I argue that this marks an historical pivot in translatorial strategy between those promoting a national poetics and those championing a world poet. As an alternative, I offer a speculative history for how Turkish poetry could have developed in translation if it had adopted the strategies of Nâzım’s translators’ instead of the anthological model. Since this dissertation constitutes part of a creative-critical project, I also enact the single-book model by translating Yücel Kayıran’s *Efsus’a Yolculuk* (Passage to Efsus). To represent this translation in my dissertation, I include my translation of his “Kâbe’ye Yolculuk,” the predecessor of his book-length poem