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
Gallium Nitride Photocathode Data
No full textA zip file containing (i) the raw data (.tif image files) for the GaN photocathode characterization that was performed in June of 2024 and (ii) the OriginLab Origin 2024 data file used to create the plots in the associated publication (https://doi.org/10.48550/arXiv.2412.17207) which, itself, includes a spreadsheet containing the MTE and QE values obtained from analyzing the raw data as well as the GaAs MTE and QE from L. Jones, 2014 that is also presented in the same publication.</p
<b><i>Pani Kowalska</i></b> - An Open-Source Textbook for Second Year Polish
No full textThis is the first chapter of a new open-source textbook for second year Polish. Please contact me for additional activities, audio files ar [email protected] - Joanna Wolf-MarszałkiewiczGraphic Design - Jamie Lin</p
Breaking Down the Monolith: Using Network Methods to Identify and Analyze Chicago’s Largest Landlords
No full textMore than a third of households in the US and more than half in Chicago, IL are renter-occupied. In recent decades, ownership of residential rental properties has become increasingly concentrated in the hands of large, sophisticated landlords. Large landlords are obscure in two senses. First, it is difficult to determine which actors own which properties. Second, the causes of landlord behavior (e.g., investment decisions and property management practices) are often complex and difficult to ascertain. In this thesis, I address both types of large landlord obscurity in Chicago, IL. First, I comprehensively identify the largest landlords and the scopes of their portfolios by linking large, administrative datasets in a network model. I identify 48 landlords who each own 1,000 or more units in MFR buildings (as of 2024). These 48 landlords collectively own 17.3% of MFR units in Chicago. After identifying these actors, I describe them, especially in terms of their business strategies and histories. Quantitatively, I analyze readily observable characteristics of parcels in their portfolios. Qualitatively, I draw from a variety of non-academic sources to conduct case studies on three landlords. To analyze business strategies, I draw from various academic sources, and I apply the popular real estate industry framework of “core,” “value-add,” and “opportunistic” investing. I demonstrate that Chicago’s large landlords are highly diverse, varying greatly in terms of firm age; the spatial distributions of their holdings; the racial and class characteristics of renters in neighborhoods where they invest; and sizes and values of owned buildings. With my case studies, I offer evidence of all three investment strategies. With this thesis, I contribute to a young but growing body of work that uses computational methods and administrative data to identify which landlords own which properties. Further, I demonstrate that Chicago’s large landlords are not monolithic. I argue that understanding landlord behaviors requires a nuanced and multi-factorial framework. By addressing both types of obscurity, this thesis bridges two largely separate bodies of literature. Such an approach allows for a strong understanding of a city’s residential rental market dynamics. Supplementary Materials for this thesis can be found at https://doi.org/10.25417/uic.29169677.v1.</p
Evaluating the Impact of GPU Frequency Tuning and Power Capping on Performance and Efficiency
No full textSUMMARY
High-Performance Computing (HPC) systems play a pivotal role in modern scientific and en-
gineering advancements. However, the increasing demand for computational power comes with
significant energy consumption challenges. This thesis investigates the impact of GPU fre-
quency tuning and power capping on performance and energy efficiency across three generations
of NVIDIA GPUs: Pascal (P100), Volta (V100), and Ampere (A100). The study leverages the
Altis Benchmark Suite to comprehensively evaluate the performance and energy behavior of
diverse workloads. Through systematic power management strategies, including power capping
and frequency tuning, we aim to identify optimal configurations that balance computational
performance and energy consumption. The results reveal that power capping is particularly
effective for compute-bound workloads, while frequency tuning provides finer-grained control
over energy efficiency, especially for architecture-specific optimizations. Additionally, the ex-
periments demonstrate that the Ampere architecture (A100) offers more precise control over
power-performance trade-offs compared to its predecessors, highlighting the evolution of power
management capabilities in modern GPUs. A novel contribution of this research is the integra-
tion of predictive modeling to estimate energy consumption under varying frequency and power
cap configurations. Three machine learning models were employed: Gaussian Process Regres-
sion (GPR), Long Short-Term Memory (LSTM) neural networks, and Random Forest (RF).
These models were trained and validated on experimental data to forecast energy consump-
tion trends, allowing for adaptive energy-aware strategies in HPC environments. The results demonstrate that GPR excels in capturing smooth, non-linear patterns, LSTM effectively han-
dles workloads with temporal dependencies, and RF provides robust performance for diverse
benchmarks. This predictive framework not only enables real-time optimization but also as-
sists in developing energy-efficient scheduling strategies for future HPC systems. Furthermore,
this work introduces a classification of benchmark behaviors into consistent and non-consistent
categories, aiding in the identification of workloads that benefit the most from specific power
management techniques. By establishing clear guidelines for frequency tuning and power cap-
ping, this thesis provides practical recommendations for optimizing energy efficiency without
compromising performance. In conclusion, this research enhances the understanding of GPU
power management techniques and introduces a predictive framework for dynamic energy op-
timization in HPC clusters. The findings pave the way for more efficient and sustainable
high-performance computing systems, enabling better resource allocation, reduced operational
costs, and improved environmental impac
Parental Perspectives on ChatGPT as an Oral Health Information Resource
No full textPurpose: Patients increasingly obtain oral health (OH) information online and from artificial intelligence (AI)
tools. As ChatGPT gains popularity, its potential for OH education remains unclear. This cross-sectional study
evaluates parents’ ChatGPT familiarity/usage and the perceived accuracy/helpfulness of responses to
pediatric-specific clinical questions.
Methods: Parents completed Qualtrics surveys between August–November 2024 in three clinical settings: a
private fee-for-service office, a private primarily Medicaid-serving office, and a university-based Medicaid-
serving clinic. Surveys collected sociodemographic and self-reported OH information. Parents also described
their experience/familiarity with ChatGPT and rated the accuracy/helpfulness of six ChatGPT responses
covering various pediatric dentistry topics (e.g., prevention, treatment, safety). Bivariate and multivariate
logistic regression were conducted to account for covariates.
Results: 234 of 302 surveys were analyzed (surveys with >75% of questions completed). Caregivers were 81%
female, 33% under 35 years, 43% White, 33% Hispanic, and 10% Black. Most were employed full-time (68%)
and had education beyond high school (71%). Children were 50% Medicaid-enrolled, 53% male, and primarily
6-to-10 years old (48%). Medicaid was associated with poorer caregiver and child self-reported OH and the
child’s likelihood of having a restorative visit (P<.001). Parents of privately insured children had greater
ChatGPT experience (P=0.037) but felt less likely to use it for obtaining OH information (P=.018). Over 80% of
participants rated ChatGPT responses as “very” or “somewhat” accurate/helpful.
Conclusions: Parents have varying experience with ChatGPT, yet parents rated its information as
accurate/helpful, regardless of SES. Further research is needed to explore the implications of these findings
for pediatric OH education
Designing Compact On-Disk Index Data Structures
No full textThe performance of a database system heavily depends on its storage engine. The design
of on-disk index data structure is critical to the storage engine performance. Many database
systems have chosen B+-trees as the on-disk index data structure. A B+-tree organizes every
internal node in a block of the I/O unit so that the reads would incur minimum I/Os. However,
writes in B+-trees are slow because a single write often causes multiple blocks to be rewritten
due to the need to maintain its structure. To enable the support for more efficient writes,
many database systems start to adopt LSM-trees in their storage engines. While an LSM-tree
achieves efficient writes by batching small updates to eliminate in-place rewrites, reads in LSM-
trees often require multiple I/Os to read the data scattered across a number of files on the disk
which makes its reads much slower than B+-trees’.
For a long time, reads and writes have been perceived to be a trade-off in designing on-
disk index data structures. It still remains a challenge to achieve efficient reads and writes at
the same time. In this thesis, we explore the idea of using compact indexes on on-disk data
structures that can achieve efficient reads and writes. In this thesis, we introduce three works
that progressively achieve more efficient index data structure designs. The first work REMIX
compactly records and replay the scanning paths to achieve more efficient scanning in LSM-
trees. The second work Disco, built on top of REMIX, achieves more efficient use of I/O for
point and range queries. Disco guarantees that a point query issues at most one I/O while
a short range query needs at most two I/Os, achieving an I/O efficiency close to a B+-tree. The third work extends the design of Disco to achieve an optimal use of I/O for on-disk data
structures
Police and Burnout: A Quantitative and Theoretical Analysis of Burnout in American Policing
No full textThe bulk of the empirical research examining burnout among police officers has focused on individual outcomes related to officer health and performance. This dissertation study suggests a more comprehensive approach leading in two directions: a quantitative analysis of three individual theories of burnout—conservation of resources (COR), job demands-resources (JD-R), and areas of worklife (AWS)—and the creation and development of an integrated theory of burnout (ITB) to understand how resources, demands, and the working environment influences officer burnout in the United States. Part I analyzes which factors are consistent with the COR, JD-R, and AWS theoretical models and are strongly related to burnout among police officers. Part II develops and assess empirically an integrated theory of burnout (ITB). This study analyzes previously collected survey data from thousands of officers representing law enforcement agencies throughout the country, in addition to surveys of their “executives” and broader departmental characteristics. The results of this study are used to explain the prevalence and incidence of burnout among police officers and to create new programs and initiatives that intend to alleviate officer’s burnout through various strategies
Cocaine Use Disorder: Anhedonia and Emotional Flattening
No full textAnhedonia, the reduced capacity to experience or pursue pleasure, is common among individuals with cocaine use disorder (CUD) and may contribute to poor treatment outcomes. This blunted emotional experience may extend beyond diminished positive affect to include reduced sensitivity to negative emotions, suggesting a broader emotional flattening. This narrowing of the emotional spectrum has been noted in other clinical populations and, if also present in individuals with CUD, may offer a key target for more effective treatment strategies.
We hypothesized that individuals with greater anhedonia would rate images more neutrally, regardless of valence or arousal, and that this emotional flattening would extend to physiological responses, including heart rate reactivity and facial muscle activity (corrugator EMG).
To test this, 53 participants (83% male) diagnosed with moderate-to-severe CUD based on DSM-5 criteria completed the Snaith-Hamilton Pleasure Scale (SHAPS), a validated 4-point measure of anhedonia, and the Emotional Picture Rating Task (EPRT). In the EPRT, participants viewed 36 emotional images (positive, neutral, and negative) from the Nencki Affective Picture System. After each 6-second image presentation, they rated the valence (-4 to +4) and arousal (0 to 9), while heart rate and corrugator EMG were continuously recorded.
Mixed-effects models revealed a significant interaction between picture type and anhedonia on valence ratings, indicating that higher anhedonia was associated with reduced differentiation in valence. A quadratic pattern emerged for arousal ratings, but anhedonia did not significantly moderate this effect. Physiologically, anhedonia did not significantly influence heart rate deceleration or acceleration, but also typical emotional heart rate patterns were not observed in the overall sample, possibly due to cocaine-related cardiac disruptions. Corrugator activity varied by picture type as expected but did not differ by anhedonia level.
These findings suggest that while anhedonia in CUD is associated with subjective emotional flattening, this is not mirrored in physiological measures. These findings highlights the potential value of interventions that increase emotional sensitivity across the full emotional spectrum. Treatments such as Acceptance and Commitment Therapy, Radically Open DBT, and interoceptive-based approaches may be particularly effective for addressing this pattern in individuals with CUD
Health Assessment of Artisanal Metal Work: Injury Patterns, Aerosol Exposures, and Biomonitoring
No full textArtisanal metal sculptors are an often independent and self-employed workforce whose contributions to the cultural landscape remain underrecognized in public health research. Despite sustained exposure to airborne particulates and acute injury hazards, these workers operate outside traditional regulatory frameworks and surveillance systems, such as those governed by OSHA, due to their non-standard employment status, with a notable lack of workers’ compensation or insurance coverage. This dissertation addresses that critical oversight through a three-aim investigation that integrates injury surveillance, environmental exposure assessment, and biomonitoring.
Aim 1 analyzed First Reports of Injury (FROI) submitted to the Illinois Workers’ Compensation Commission from 2019 to 2023. Because sculptors lack a dedicated occupational classification, a surrogate group was identified using text-mining and recoding methods. This group demonstrated higher rates of injury claims and longer durations of uncompensated lost workdays compared to machinists and metal processors. These findings highlight systemic under-recognition of sculptors in injury databases and call for more inclusive surveillance strategies.
Aim 2 characterized environmental exposures in seven metal sculpture studios across the Chicago metropolitan area. Real-time aerosol monitoring and personal air sampling revealed short bursts of intense particulate emissions during active tasks, particularly in spaces with poor ventilation. Although base metals varied, the most frequently detected airborne metals-such as aluminum, manganese, copper, chromium, cobalt, and nickel-were often linked to auxiliary materials like abrasives and welding consumables rather than the sculpted workpieces themselves. These findings underscore that exposure risks are shaped more by materials and task processes than by the core metal medium.
Aim 3 assessed internal exposure among 12 sculptors using matched pre- and post-work samples across four biological matrices: blood, urine, saliva, and exhaled breath condensate (EBC). Blood samples yielded the most consistent quantification, particularly for manganese and chromium, and correlated with personal air concentrations. Saliva and EBC also showed promise as non-invasive alternatives for the biomonitoring of specific metals, offering potentially feasible alternatives to field-based surveillance when phlebotomy is impractical
Design and Development of Novel Atomic Layer Deposition Equipment & Processes for Biomedical Applications
No full textUltra-thin films grown using Atomic Layer Deposition (ALD) have gained a lot of ground in recent years. This technology, which is still relatively new, has created a benchmark for depositing pure, conformal, pin-hole free films on almost any substrate with atomic scale precision. One of the major applications of this deposition technique has been in the biomedical implant industry where ALD can be used to functionalize many commonly used biomaterials to enhance their surface properties. These biomaterials have special applications but due to their organic nature, they are not receptive to cell proliferation. ALD has become indispensable mostly because of its ability to coat organic substrates with enough material to induce cell proliferation but not enough to induce a cytotoxic response. During research into low temperature ALD reactions, work was done to produce ultra-thin films of Titanium dioxide (TiO2) or titania, magnesium oxide (MgO), titanium oxynitride(TiOxNy) and silver (Ag) metal. Each metal or oxide explored provided some enhancement to the underlying substrate in terms of hydrophobicity, cell response and osteointegration. The as-deposited films were characterized by spectroscopic ellipsometry (SE) for thickness and refractive index. Films were also characterized using X-ray Photoelectron Spectroscopy (XPS) to study its chemical composition. Furthermore, density and absolute thickness of as-deposited films were also measured using X-Ray Reflection (XRR). Mechanical testing included measurement of the tensile strength of the material before and after deposition. Leaching of the coated material in Phosphate Buffer Solution (PBS) was also studied using Inductively Coupled Plasma- Quadrupole Mass Spectrometry (ICP-QMS). Through these studies, low temperature reactions were demonstrated on substrates such as polymethylmethacrylate (PMMA), bone particles and collagen membrane that are temperature sensitive. A reactor for bilayer ALD of metal oxide/nitride is also presented. With this reactor, a novel approach was used to control the reactor using a custom code written in Python and executed using a commonly available single board computer (SBC) and microcontrollers. This led to the design being more compact and robust. This reactor was used to deposit Titanium oxynitride and this reaction was studied using a QMS attached to the downstream of the reactor to get insights into the ALD reaction
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! 👇