University of Illinois at Chicago
University of Illinois at Chicago: UIC INDIGO (INtellectual property in DIGital form available online in an Open environment)Not a member yet
21439 research outputs found
Sort by
Contrasting Perceived Barriers to Recreation for People with Disabilities in Maywood Park District
No full textRemoving accessibility barriers to special recreation in community park districts, particularly among kids and youth with disabilities has been identified as an important need. The purpose of this study was to identify barriers in the Maywood Park District policies, practices, and spaces impacting the growing Latino population. Initial community needs assessments indicated that the Latino population was not represented in the park district needs assessments. The research methodology included quantitative measures (the Americans with Disabilities Act Accessibility Guidelines checklist (ADAAG) and the Inclusivity Assessment Tool (Inclusion U) to identify physical barriers and policies in the Park District. I also conducted individual interviews with staff members and 2 focus group meetings with community members, one of them with Latino community members. The instruments identified several barriers which were confirmed and expanded by the staff interviews and the focus group data. The Spanish speaking focus group reported cultural factors as significant barriers to community access for people with disabilities in the Maywood Park District. The study findings indicated that public communication and awareness were the greatest barriers. I hope that the results of the study will be used to locate and prioritize barrier removal opportunities for the Maywood Park District
Deep Joint Denoising and Compression for Satellite Images
No full textEmerging imaging systems must handle large amounts of noisy data under strict resource constraints. Instead of denoising and compressing in two separate steps, this work integrates both tasks into a single framework based on the Mean Scale Hyperprior architecture. By incorporating denoising directly in the compression process, the model dedicates fewer bits to encoding noise and focuses on preserving essential image details.
Two training modalities are explored: one where the model is trained to reconstruct noisy inputs, and another where it learns to produce clean outputs from noisy inputs. Experiments using SEN12MS satellite images and a custom noise model show that models trained to generate clean outputs achieve higher-quality reconstructions and better rate-distortion performance. This integrated approach not only reduces computational and hardware complexity but also improves bandwidth efficiency, offering a compelling alternative to traditional two-step pipelines
Towards an Informative Recommender System
No full textRecommender systems play a critical role in modern digital platforms, yet their effectiveness depends on how well they extract and utilize implicit information from available data. This dissertation explores informative learning—a paradigm that enhances recommendation quality by extracting, refining, and transferring knowledge from structured and unstructured data.
We introduce a series of informative learning techniques that advance market adaptation, confidence calibration, collaborative-semantic alignment, knowledge extraction, and self-optimizing ranking. Specifically, we propose a transferable attention mechanism for adapting recommendation models across diverse geographic markets without requiring additional training data. A confidence-aware fine-tuning framework integrates conformal prediction to quantify uncertainty in recommendations. To bridge collaborative and semantic signals, we develop a method that improves performance in both warm-start and cold-start scenarios by leveraging textual information. We further explore how large language models (LLMs) can generate structured representations from unstructured game text to enhance personalization and content integrity. Finally, we present Auto-Guided Prompt Refinement (AGP), a dynamic framework that refines user profiles based on ranking feedback without manual prompt engineering.
Extensive experiments on multiple real-world datasets validate the effectiveness of the proposed techniques. Collectively, these contributions advance the development of adaptive, confidence-aware, and knowledge-enriched recommendation systems
Mechanisms of Western-Diet Induced Endothelial Stiffening: Endothelial CD36 and Long-Chain Fatty Acids
No full textWestern diet (WD) refers to a diet high in saturated fat and refined carbohydrates. WD is known to be detrimental to cardiovascular health. These studies present novel insights into the mechanism undergirding WD-associated cardiovascular dysfunction: stiffening (an increase in the elastic modulus) of the aortic endothelial layer associates with barrier disruption, and both are mediated by endothelial scavenger lipid receptor CD36. Through the use of an inducible, endothelial-specific CD36 knockdown mouse model, we discovered that loss of endothelial CD36 prevents endothelial stiffening and minimizes barrier leakage. We discovered that WD-induced endothelial stiffening in vivo is mimicked by exposure to saturated LCFA palmitic acid in vitro, a phenomenon that is dependent on CD36. Activation of CD36 by non-lipid ligands, however, does not elicit stiffening. To elucidate how lipids elicit stiffening, we focused on RhoA and its association with RhoGDI1, which maintains RhoA inactive. We show that RhoA and RhoGDI1 are required for palmitic acid-induced stiffening. These findings point to RhoGDI1 as a target for the development of studies into WD-induced cardiovascular dysfunction
Pleasures and Powers of Affective Disorientation: Iconoclastic Ambivalence and Polyvalence in the NDW
No full textThis dissertation analyzes the rich affective dynamics characterizing the music of the Neue Deutsche Welle (NDW) and its ambivalent critical reception, with a particular focus on gender representation. It argues that NDW music expresses the conflicted structures of feeling (Williams) that characterize the era, so that the works and praxes considered embody both empathic community-centeredness and a tragic adoption of market-based commodification.
The NDW was a cultural phenomenon that developed between 1978 and 1983 from an intimate network of local, subcultural scenes into the defining national pop-musical genre in the Federal Republic of Germany. The dissertation focuses on the initial recordings of three key women-led bands, whose successes indicate the arc of the larger movement: Hans-A-Plast, Ideal and Nena. Through a series of historically situated and theoretically inflected close readings, the project reveals how these bands’ songs incorporate ambiguity in a novel way, specifically as a reparative, representational scheme. Reparative (Sedgwick) suggests a communal approach to healing psychic wounds caused by stresses inherent in post-fascist, post-1968 German society. And the songs deploy ambiguities representationally in strategic acknowledgement of the complexity of life as experienced by the musicians and their communities.
That the NDW was home to an increased number of women and other underrepresented groups in pop music, and that these artists’ work was explicitly collaborative, lends both the new aesthetic and the hostile critical reception additional significance. Mapping these diverse affective relationships enables the dissertation to articulate the structures of feeling captured by the NDW and to demonstrate the deeply political heart of this music, especially the frank descriptions of gendered experience therein. Hence, the project refutes critics’ accusations of the music’s narcissistic superficiality or meaningless frivolity. Instead, the project shows, the NDW precisely eschews didactic sloganeering in favor of semantic abundance. Simultaneously however, the artists adopted market-based models of commodification that undermined their egalitarian intentions. This also demonstrates a structure of feeling of the era, and the concluding chapter explores this fundamental contradiction
MEMS Hybrid Electronic Integration: From Micro-Robotic Actuation to Control and Monitoring of MEMS Sensor
No full textThis work presents two hardware implementations for controlling MicroElectroMechanical
Systems (MEMS) in mechatronic systems across different scales.
The first project focuses on advancing micro-robotic swarm control through improvements to
Field Programmable (FP) MicroStressBots. These micro-robots rely on electrostatic actuation
and utilize Untethered Scratch Drive Actuation (USDA) for translational motion. Individual
selectivity and rotational control are achieved through an electrostatically actuated steering
arm, which is programmed via stress-engineering of the arm. While several theoretical control
strategies for micro-robotic swarms have been proposed, they often neglect the non-idealities
associated with programming constraints. Although the command complexity remains O(c), in
practice, only a finite number of micro-robots can be operated simultaneously without address
conflicts.
To address this limitation, this work introduces an expanded steering arm design incorporating
MEMS relay logic to electrically isolate the steering arm. The addition of Normally
Open (NO) and Normally Closed (NC) cantilevers enables power delivery only when a valid
addressing sequence is provided. Programming of these relay logic cells is achieved through a
new stress-engineering method using localized shadow masks fabricated by two-photon polymerization
(2PP) micro-scale 3D printing. These 3D-printed masks feature geometries designed
for simple release and precise placement over the MEMS relay logic cells. This development
represents a significant step toward practical micro-robotic swarm control.
In parallel, two exploratory studies were conducted to enhance electrostatic micro-robotic
systems. The first investigated the use of high-permittivity (high-k) dielectrics to strengthen
electrode fields and potentially reduce operating voltages for FP MicroStressBots. The second
explored the fabrication of electrostatic devices by sputter-coating 2PP 3D-printed structures
with metal.
The second project addresses the need for monitoring respirable particulate matter (PM)
smaller than 4 μm, which can remain suspended in air for extended periods and is prevalent
in environments such as coal mines and diesel exhaust. These particles are linked to respiratory
illnesses, including Coal Workers’ Pneumoconiosis (CWP), lung cancer, and silicosis, with
disease severity inversely correlated with particle size.
This work builds upon the Wearable Respirable Dust Monitor (WEARDM), a dual MEMS
gravimetric sensor platform for high-sensitivity PM detection. The original WEARDM used a
Quartz Crystal Microbalance (QCM) to detect particles between 4 μm and 1.3 μm, and a Film
Bulk Acoustic Resonator (FBAR) for particles below 1.3 μm. These sensors were previously
characterized using benchtop equipment to determine sensitivity and saturation limits. To
advance toward a wearable, self-contained solution, this work integrates the system onto a single
printed circuit board (PCB) containing a microcontroller, a radio-frequency (RF) conditioning
circuit, and a five-output power supply to power and monitor the sensors. This marks the next
logical step toward a fully wearable respirable dust monitoring device
A conceptual model for co‐developing a culturally tailored intervention for Latina immigrant caregivers of children with disabilities
No full textThe growing diversity of the U.S. population, partly due to immigration, has called attention to scholars and practitioners to attend to immigrants' cultural beliefs, values, and ways of doing when designing interventions to promote health and wellbeing. In this paper, we propose a contextual and dynamic model for co-developing a culturally tailored intervention with the community to advance equity and empowerment of Latinx immigrant caregivers of children with intellectual and developmental disabilities (IDD). Grounded in the literature and voices of the community, the proposed model includes six interactive dimensions (LARREDS) that guided the development of the PODER Familiar intervention described here. These include language and linguistic preferences; accessibility factors; reflecting the group's values, ways of thinking and doing; reflecting generational differences; dimensions of delivery and learning style; and the social, ecological, and cultural environment. Informed by principles of family engagement, the model also includes eight strategies for engaging caregivers throughout the intervention. The conceptual model was co-developed with promotoras who also provided input on the PODER Familiar intervention. While describing the model in action, we highlight the voices of the promotoras. The implications of culturally tailored interventions and the application of the model to designing interventions for other migrant populations are discussed.</p
Angle-Resolved Photoemission Studies on the Many-Body Effects of High-Temperature Superconductor: Bi-2212
No full textSpin fluctuations are defined as dynamic variations of spin density in both space and time. They are widely considered to play a essential role in the pairing mechanism of high temperature superconductors. Unconventional superconductors such as cuprates are believed to exhibit spin-fluctuation-mediated pairing, as opposed to conventional superconductors where phonons mediate Cooper pairing. Dynamic spin susceptibility can characterize these fluctuations. It describes the magnetic response of the system to perturbations. The imaginary part of this function captures the full spectrum of spin excitations in both the normal and superconducting states. It can be typically measured via inelastic neutron scattering (INS). However, such measurements for Bi-2212 are rare because of small scattering cross-sections and difficulties in growing sufficiently large crystals. In this work, we developed a stable photon-energy helium lamp–based ARPES (angle-resolved photoemission spectroscopy) system with high energy resolution to extract detailed single-particle spectral functions in Bi-2212. Using these spectral functions, we reconstructed both the non-interacting and interacting spin susceptibilities in the normal and superconducting states. The ARPES-derived susceptibilities reveal a broad continuum of spin excitations in the normal state and a sharp spin resonance emerging in the superconducting state due to the collective spin-1 excitation. These results reflect the underlying correlated electron dynamics in Bi-2212 and support the picture of spin fluctuations as a key mediator of high temperature superconductors
Guardianship Experiences: People With Intellectual and Developmental Disabilities and Their Caregivers
No full textApproximately 1.3 million Americans are currently under guardianships or conservatorships of some form, meaning 1.3 million family members, friends, and state-appointed guardians also must navigate being guardians. Therefore, the legal processes and perspectives of guardians and individuals under guardianship are valuable in research and policy development. As guardianship data is limited, it is necessary to better understand the experiences and perspectives of both guardians and those under guardianship to better advocate for guardianship reform. Five focus groups examined the perspectives of fifteen caregivers and thirteen individuals with intellectual and developmental disabilities on how they accessed information on guardianship, and on their internal experiences with guardianship. Findings on information access included the perspectives on “Roles of a Guardian,” “Guardianship and Supported Decision-Making Processes,” and “Guardianship as a System.” Findings on the internal experiences with guardianship included perspectives on “Benefits of Guardianship,” “Barriers of Guardianship,” and “Emotional Outcomes.” A review of guardianship history and alternatives, both from the federal and state perspectives, gave context and framing to the reported experiences and perspectives
Optimism and Robustness: Learning From Structured and Semi-Random Inputs
No full textTraditionally, algorithms have been studied under two regimes: worst-case analysis, which makes no assumptions about the input, and average-case analysis, which assumes that inputs are drawn from a certain distribution. However, real-world inputs rarely conform to either of these extremes. A new paradigm known as “beyond worst-case analysis” seeks to bridge the gap between the two. In this thesis, we study several problems and their algorithms under appropriate beyond worst-case models, aiming to provide more realistic and practically relevant performance guarantees.
In the first part of this thesis, we focus on improving algorithm performance on non-worst-case inputs (structured inputs). In Chapter 2, we study the Boolean satisfiability problem (SAT) in the framework of learning-augmented algorithms, where the problem instance is provided with a prediction that contains partial information of an optimal
solution. We study both the decision and optimization problem of SAT under two forms of predictions, namely the subset advice and the label advice. In Chapter 3, we study non-center-based clustering under Bilu-Linial stability assumptions, which assumes that the problem instance has a unique optimal solution that stays unchanged under small
perturbation of the input. We focus on the minimizing sum-of-radii (MSR) and minimizing sum-of-diameters (MSD) objectives, and provide polynomial time solutions under stability assumptions.
In the second part of this thesis, we focus on enhancing algorithm robustness to contamination in average-case inputs (semi-random inputs). In the context of low-rank matrix recovery problems, this means a monotone adversary can add arbitrary data from the distribution to break the necessary regularity conditions satisfied by fully random inputs. In
Chapter 4, we study the matrix completion problem, whose goal is to recover a ground-truth matrix from incomplete and noisy observations of its entries. In Chapter 5, we study the matrix sensing problem, where the goal is to recover the ground-truth matrix based on linear measurements from a given set of sensing matrices
Data sources:
Access Repository Dashboard
Do you manage Open Research Online? Become a CORE Member to access insider analytics, issue reports and manage access to outputs from your repository in the CORE Repository Dashboard! 👇