University of Illinois Urbana-Champaign
Illinois Digital Environment for Access to Learning and Scholarship RepositoryNot a member yet
123813 research outputs found
Sort by
Hercules: A compiler for productive programming of heterogeneous systems
Get PDFSubmission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2025-10-19 without embargo termsThe student, Russel Arbore, accepted the attached license on 2025-04-15 at 20:55.The student, Russel Arbore, submitted this Thesis for approval on 2025-04-15 at 21:01.This Thesis was approved for publication on 2025-04-16 at 07:41.DSpace SAF Submission Ingestion Package generated from Vireo submission #21788 on 2025-10-19 at 18:09:24Modern computing systems increasingly rely on composing heterogeneous devices to improve performance and efficiency. Programming these systems is often unproductive: algorithm implementations must be coupled to system-specific logic, including device-specific optimizations, partitioning, and inter-device communication and synchronization, which requires developing different programs for different system configurations. We propose the Juno language, which represents general purpose applications in an imperative form that can be transformed into parallel, optimized, system-specific code using an expressive and granular imperative scheduling language. We also introduce the Hercules compiler, which uses a novel intermediate representation to represent general and device-specific parallel code in a manner that is easy to analyze and manipulate using schedules. Our system achieves competitive performance with hand-optimized device-specific code (geomean speedups of 1.25x and 1.48x on the CPU and GPU) and significantly outperforms a prior general purpose heterogeneous programming system (geomean speedups of 9.31x and 16.18x on the CPU and GPU)
Event-based knowledge editing for deterministic knowledge propagation in large language models
Get PDFSubmission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2025-10-19 without embargo termsThe student, Jiateng Liu, accepted the attached license on 2025-04-17 at 10:21.The student, Jiateng Liu, submitted this Thesis for approval on 2025-04-17 at 10:21.This Thesis was approved for publication on 2025-04-17 at 13:26.DSpace SAF Submission Ingestion Package generated from Vireo submission #21819 on 2025-10-19 at 18:09:31The dynamic nature of real-world information necessitates knowledge editing (KE) in large language models (LLMs). This edited knowledge should propagate and facilitate the deduction of new information based on existing model knowledge. We define the existing related knowledge in a LLM serving as the origination of knowledge propagation as ``deduction anchors''. However, most of current KE approaches only operate on (subject, relation, object) triples. Both theoretically and empirically, we observe that this simplified setting often leads to uncertainty when determining the deduction anchors, causing low confidence in their responses. To mitigate this issue, we propose a novel task of event-based knowledge editing that pairs facts with event descriptions. This task manifests both as a closer simulation of real-world editing scenarios and a more logically sound setting, implicitly defining the deduction anchor and enabling LLMs to propagate knowledge confidently. We curate a new benchmark dataset evedit derived from the CounterFact dataset and validate its superiority in improving model confidence. Moreover, as we observe that the event-based setting is notably challenging for existing approaches, we propose a novel approach Self-Edit that showcases stronger performance, achieving 55.6\% consistency improvement while maintaining the naturalness of generation
Safety, security, and rehabilitation? An analysis of publication censorship practices and previously imprisoned individuals’ experiences in Illinois
Get PDFSubmission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2025-10-19 without embargo termsThe student, Carileigh Jones, accepted the attached license on 2025-04-21 at 20:32.The student, Carileigh Jones, submitted this Dissertation for approval on 2025-04-21 at 20:38.This Dissertation was approved for publication on 2025-04-25 at 13:18.DSpace SAF Submission Ingestion Package generated from Vireo submission #21860 on 2025-10-19 at 18:09:35Scholarly discourses surrounding publication censorship policies and practices have virtually ignored the voices of the imprisoned. While these discourses help us to understand, and critique, the idea that book censorship promotes rehabilitation, safety, and security, they offer top-down analysis, which take for granted the viewpoints and experiences of those who have been incarcerated. Departing from this line of thinking, my dissertation theorizes the censorship of reading material within U.S. prisons as a practice of carcerality, which not only defines and regulates the limits of what can be known and expressed, but may also be experienced by prisoners as control and surveillance of both the mind and body (Friedman, 2021). Located at the intersections of sociology, critical criminology, and carceral studies, this dissertation analyzes interviews, court case documents, and policy documents to place explanations of the purpose and goals of censorship from the perspective of prison officials, in conversation with the perceptions and experiences of those who have been incarcerated. I argue that centering the voices of the imprisoned when implementing and assessing censorship guidelines is crucial if the purpose of books, and prison reform more generally, is to improve the lives and experiences of the imprisoned. Taking their perspectives into account could 1) enhance policy work related to book censorship practices, 2) promote a level of agency in the lives of the imprisoned, and 3) reshape the ways we understand discipline and rehabilitation within and beyond prison walls
Reciprocal influences of caregiver language facilitation strategies and child characteristics on caregiver-child interactions
Get PDFSubmission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2025-10-19 without embargo termsThe student, Emily Harrington, accepted the attached license on 2025-04-23 at 14:12.The student, Emily Harrington, submitted this Dissertation for approval on 2025-04-23 at 15:49.This Dissertation was approved for publication on 2025-04-24 at 15:14.DSpace SAF Submission Ingestion Package generated from Vireo submission #21921 on 2025-10-19 at 18:10:51Caregiver-implemented interventions have been shown to have a positive impact on child language outcomes. While caregiver language facilitation strategies play a major role in this, an unanswered question is what specific strategies are tied to improvements in child language and what mechanisms drive this change. As a first step in understanding this, the current study examined the baseline caregiver-child interactions (CCXs) of caregivers and their 30-month-old children at risk for developmental language disorder (DLD) prior to an intensive, research-based intervention. By bringing together the transactional model of development and empirical evidence on language facilitation strategies, I examined associations between child factors and caregiver use of language facilitation strategies and, in turn, how these strategies influenced children’s real time communication. Dyads (N = 105) were participants in the Enhanced Milieu Teaching-Sentence Focused clinical trial. Behavioral assessments, caregiver report measures, and CCXs were collected when children were approximately 30 months of age. CCXs were transcribed and coded for four caregiver language facilitation strategies and child communication variables. LASSO regression and post-selection inference indicated that caregiver report measures of children’s vocabulary and temper loss behaviors were significantly associated with the frequency of caregiver strategy use within the CCXs. LASSO regression and post-selection inference also indicated that child communication within the CCXs was associated with both the frequency and type of strategies caregivers used. Namely, children’s overall communication acts were significantly associated with caregiver use of vocabulary strategies, while children’s vocabulary diversity was significantly associated with caregiver use of sentence strategies. Mixed-effects logistic regression demonstrated that all four caregiver strategies significantly increased children’s likelihood of spontaneous imitation on their next turn. These findings reveal that, prior to intervention, caregivers and children reciprocally shape their language environment. This supports a transactional perspective and warrants further consideration of reciprocal influences when assessing the impact of caregiver-implemented interventions
A standard operating procedure for accurate condensation heat transfer measurements
No full textSubmission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2025-10-19 without embargo termsThe student, Ghassan Arissi, accepted the attached license on 2025-04-25 at 10:22.The student, Ghassan Arissi, submitted this Thesis for approval on 2025-04-25 at 11:16.This Thesis was approved for publication on 2025-05-06 at 16:57.DSpace SAF Submission Ingestion Package generated from Vireo submission #21977 on 2025-10-19 at 18:11:01The condensation of steam is an important phase change process to many industries, its high complexity however makes it one of the hardest processes to quantify. The experimental data collected for dropwise condensation in particular, has been very scattered and that has mainly been attributed to the presence of trace amounts of non-condensable gases (NCGs) in the steam. It is the objective of this thesis to develop a standard operating procedure that can easily be implemented in condensation chambers around the world to enable accurate heat transfer measurements. Here the use of a vent as a means to deal with NCGs is explored. Through experimental work, analytical modeling, and CFD simulations new insights are made on the role of the vent. Finally, guidelines are given to help in properly guiding the use of a vent to limit NCG effects without otherwise influencing the heat transfer data collected
The (im)possible queer south: Exploring archival description of southern LGBTQ+ materials across the United States
Get PDFSubmission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2025-10-19 without embargo termsThe student, Evan Allgood, accepted the attached license on 2025-04-28 at 17:53.The student, Evan Allgood, submitted this Thesis for approval on 2025-04-28 at 17:57.This Thesis was approved for publication on 2025-04-29 at 11:13.DSpace SAF Submission Ingestion Package generated from Vireo submission #22052 on 2025-10-19 at 18:11:09In collaboration with Invisible Histories, a 501(c)3, community-based archive based in Birmingham, Alabama that aims to preserve queer history in the American South, this thesis aims to explore good practices for uncovering southern LGBTQ+ collections within a variety of repositories. Aligning with Invisible Histories’ work to preserve southern LGBTQ+ history through acting as an intermediary between institutions, organizations, and everyday people, this project supports their mission to make southern LGBTQ+ history more accessible to researchers at any level, from novice to expert. Given their aim to ensure researchers can find the materials they seek as independently as possible—honoring the potentially complex affective experiences involved in accessing southern LGBTQ+ history—this thesis aims to build a new set of resources to aid these researchers. This project explores finding aids for collections related to southern LGBTQ+ individuals, organizations, and events to determine 1) what language is used, 2) where this language appears, and 3) how researchers can use this knowledge to improve their search practices. Beginning with a review of archival anxiety, minority stress, symbolic annihilation, and representational belonging to set the stakes, the paper then analyses 312 finding aids across 42 United States archival repositories all related to LGBTQ+ southern identity and experience to surface patterns in language and structure. The paper concludes by providing recommendations for researchers seeking out this variety of material and situating us in the present political and social climate of the United States with the aim that this so-called invisible history can be made visible
¡Si se pudo! Exploring Mexican American college student persistence stories during COVID-19
Get PDFSubmission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2025-10-19 without embargo termsThe student, Arianna Holterman, accepted the attached license on 2025-04-28 at 21:26.The student, Arianna Holterman, submitted this Dissertation for approval on 2025-04-28 at 21:39.This Dissertation was approved for publication on 2025-04-29 at 15:56.DSpace SAF Submission Ingestion Package generated from Vireo submission #22057 on 2025-10-19 at 18:11:10The primary purpose of this dissertation is to explore how Mexican American students at the University of Illinois Urbana-Champaign (Illinois) persisted in Higher Education during the COVID-19 pandemic. This descriptive qualitative study collected oral history interviews to serve as primary resources for the Student Life and Culture Archives at Illinois preserving what it was like to complete high school and succeed at the University of Illinois Urbana-Champaign during the pandemic. Particular attention is spent on the sources of persistence, the protective factors, that ultimately allowed these specific students to persist despite a large-scale emergency. The findings demonstrate that these students persisted by tapping into their existing cultural wealth assets, using individual agency, and connecting to specific Illinois resources, such as need based financial aid and identity specific spaces, groups, or resources, to mitigate the negative impacts of their experience. These stories may also be critical in understanding Mexican American student retention during not only the COVID-19 pandemic, but other personal or large-scale crisis or emergencies that may acutely impact this population due to existing social inequities
Development of lignin-PEO interlocked composite phase change materials for thermal stability and structural retention in battery thermal management system
Get PDFSubmission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2025-10-19 without embargo termsThe student, Jun Sang Park, accepted the attached license on 2025-04-29 at 12:56.The student, Jun Sang Park, submitted this Thesis for approval on 2025-04-29 at 13:15.This Thesis was approved for publication on 2025-05-07 at 15:23.DSpace SAF Submission Ingestion Package generated from Vireo submission #22087 on 2025-10-19 at 18:11:18Recently, with the rapid development of high-performance technologies such as electric vehicle batteries, data centers, and artificial intelligence systems, technologies that effectively control excessive heat generated by these systems are becoming important. Phase Change Materials (PCMs), which is attracting attention as a representative thermal management material, absorbs and releases heat by utilizing high latent heat properties, but existing PCM materials have limitations such as shape instability, liquid leakage, and environmental problems in the phase change process. In this research, a new PCM composite material was developed by mixing polyethylene oxide (PEO) and lignin, a natural polymer. The interlocked structure between PEO and lignin formed through a chemical crosslinking reaction and melt mixing process has excellent thermal stability and high shape maintenance properties, and phase change was suppressed even at high temperatures above 115°C to solve the liquid leakage problem. As a result of the experiment, it was found that the produced PCMs maintains a stable shape and excellent thermal control performance even in a repetitive thermal cycle environment. Therefore, the lignin-PEO-based PCMs proposed in this study is expected to be used as an effective thermal control material that simultaneously satisfies stability, durability, and sustainability in various high load environments including the Battery Thermal Management System (BTMS)
Neural network based method for solving SMT problems
Get PDFSubmission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2025-10-19 without embargo termsThe student, Keyu Lu, accepted the attached license on 2025-05-01 at 20:25.The student, Keyu Lu, submitted this Thesis for approval on 2025-05-01 at 20:39.This Thesis was approved for publication on 2025-05-05 at 12:28.DSpace SAF Submission Ingestion Package generated from Vireo submission #22166 on 2025-10-19 at 18:11:31Satisfiability Modulo Theories (SMT) over nonlinear real arithmetic (NRA) represents a fundamental yet notoriously difficult problem class in formal verification and symbolic reasoning. Traditional SMT solvers struggle with the scalability and decidability of QF_NRA problems due to their intrinsic nonlinearity. In this work, we propose a novel reduction-based framework that translates SMT problems defined in the SMT-LIB2 format into equivalent neural network verification problems specified in the VNN-LIB format. By constructing tailored neural networks that capture the semantics of the original constraints, we leverage powerful neural network verifiers—specifically, the α,β-CROWN solver—to determine the satisfiability of the original NRA formulas. This transformation enables the application of recent advances in neural network verification to a broader class of symbolic problems. Our approach bridges the gap between symbolic logic reasoning and neural verification, potentially unlocking new paths for scalable and parallelizable SMT solving. We demonstrate the soundness and feasibility of the method through illustrative case studies and analyze its performance in terms of accuracy and approximation fidelity
Zero-Shot Low Light Image Enhancement with Diffusion Prior
Get PDFSubmission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2025-10-19 without embargo termsThe student, Joshua Cho, accepted the attached license on 2025-05-04 at 14:40.The student, Joshua Cho, submitted this Thesis for approval on 2025-05-04 at 14:49.This Thesis was approved for publication on 2025-05-07 at 13:06.DSpace SAF Submission Ingestion Package generated from Vireo submission #22202 on 2025-10-19 at 18:12:30In this work, we present a simple yet highly effective "free lunch" solution for low-light image enhancement (LLIE), which aims to restore low-light images as if acquired in well-illuminated environments. Our method necessitates no optimization, training, fine-tuning, text conditioning, or hyperparameter adjustments, yet it consistently reconstructs low-light images with superior fidelity. Specifically, we leverage a pre-trained text-to-image diffusion prior, learned from training on a large collection of natural images, and the features present in the model itself to guide the inference, in contrast to existing methods that depend on customized constraints. Comprehensive quantitative evaluations demonstrate that our approach outperforms SOTA methods on established datasets, while qualitative analyses indicate enhanced color accuracy and the rectification of subtle chromatic deviations. Furthermore, additional experiments reveal that our method, without any modifications, achieves SOTA-comparable performance in the auto white balance (AWB) task