Digital Commons Network^™

Robust Asynchronous Federated Learning With Time-Weighted And Stale Model Aggregation, Yinbin Miao, Ziteng Liu, Xinghua Li, Meng Li, Hongwei Li, Kim-Kwang Raymond Choo, Robert H. Deng

Research Collection School Of Computing and Information Systems

Federated Learning (FL) ensures collaborative learning among multiple clients while maintaining data locally. However, the traditional synchronous FL solutions have lower accuracy and require more communication time in scenarios where most devices drop out during learning. Therefore, we propose an Asynchronous Federated Learning (AsyFL) scheme using time-weighted and stale model aggregation, which effectively solves the problem of poor model performance due to the heterogeneity of devices. Then, we integrate Symmetric Homomorphic Encryption (SHE) into AsyFL to propose Asynchronous Privacy-Preserving Federated Learning (Asy-PPFL), which protects the privacy of clients and achieves lightweight computing. Privacy analysis shows that Asy-PPFL is indistinguishable under …

Empathyear : An Open-Source Avatar Multimodal Empathetic Chatbot, Hao Fei, Han Zhang, Bin Wang, Lizi Liao, Qian Liu, Erik Cambria

Research Collection School Of Computing and Information Systems

This paper introduces EmpathyEar, a pioneering open-source, avatar-based multimodal empathetic chatbot, to fill the gap in traditional text-only empathetic response generation (ERG) systems. Leveraging the advancements of a large language model, combined with multimodal encoders and generators, EmpathyEar supports user inputs in any combination of text, sound, and vision, and produces multimodal empathetic responses, offering users, not just textual responses but also digital avatars with talking faces and synchronized speeches. A series of emotion-aware instruction-tuning is performed for comprehensive emotional understanding and generation capabilities. In this way, EmpathyEar provides users with responses that achieve a deeper emotional resonance, closely emulating …

Simc 2.0: Improved Secure Ml Inference Against Malicious Clients, Guowen Xu, Xingshuo Han, Tianwei Zhang, Shengmin Xu, Jianting Ning, Xinyi Huang, Hongwei Li, Deng, Robert H.

Research Collection School Of Computing and Information Systems

In this paper, we study the problem of secure ML inference against a malicious client and a semi-trusted server such that the client only learns the inference output while the server learns nothing. This problem is first formulated by Lehmkuhl et al. with a solution (MUSE, Usenix Security’21), whose performance is then substantially improved by Chandran et al.'s work (SIMC, USENIX Security’22). However, there still exists a nontrivial gap in these efforts towards practicality, giving the challenges of overhead reduction and secure inference acceleration in an all-round way. Based on this, we propose SIMC 2.0, which complies with the underlying …

A Multimodal Foundation Agent For Financial Trading : Tool-Augmented, Diversified, And Generalist, Wentao Zhang, Lingxuan Zhao, Haochong Xia, Shuo Sun, Jiaze Sun, Molei Qin, Xinyi Li, Yuqing Zhao, Yilei Zhao, Xinyu Cai, Longtao Zheng, Xinrun Wang, Bo An

Research Collection School Of Computing and Information Systems

Financial trading is a crucial component of the markets, informed by a multimodal information landscape encompassing news, prices, and Kline charts, and encompasses diverse tasks such as quantitative trading and high-frequency trading with various assets. While advanced AI techniques like deep learning and reinforcement learning are extensively utilized in finance, their application in financial trading tasks often faces challenges due to inadequate handling of multimodal data and limited generalizability across various tasks. To address these challenges, we present FinAgent, a multimodal foundational agent with tool augmentation for financial trading. FinAgent's market intelligence module processes a diverse range of data-numerical, textual, …

Topic Modeling On Document Networks With Dirichlet Optimal Transport Barycenter, Ce Zhang, Hady Wirawan Lauw

Research Collection School Of Computing and Information Systems

Text documents are often interconnected in a network structure, e.g., academic papers via citations, Web pages via hyperlinks. On the one hand, though Graph Neural Networks (GNNs) have shown promising ability to derive effective embeddings for such networked documents, they do not assume a latent topic structure and result in uninterpretable embeddings. On the other hand, topic models can infer semantically interpretable topic distributions for documents by associating each topic with a group of understandable key words. However, most topic models mainly focus on plain text within documents and fail to leverage network structure across documents. Network connectivity reveals topic …

Neural Network Semantic Backdoor Detection And Mitigation: A Causality-Based Approach, Bing Sun, Jun Sun, Wayne Koh, Jie Shi

Research Collection School Of Computing and Information Systems

Different from ordinary backdoors in neural networks which are introduced with artificial triggers (e.g., certain specific patch) and/or by tampering the samples, semantic backdoors are introduced by simply manipulating the semantic, e.g., by labeling green cars as frogs in the training set. By focusing on samples with rare semantic features (such as green cars), the accuracy of the model is often minimally affected. Since the attacker is not required to modify the input sample during training nor inference time, semantic backdoors are challenging to detect and remove. Existing backdoor detection and mitigation techniques are shown to be ineffective with respect …

Clamber: A Benchmark Of Identifying And Clarifying Ambiguous Information Needs In Large Language Models, Tong Zhang, Peixin Qin, Yang Deng, Chen Huang, Wenqiang Lei, Junhong Liu, Dingnan Jin, Hongru Liang, Tat-Seng Chua

Research Collection School Of Computing and Information Systems

Large language models (LLMs) are increasingly used to meet user information needs, but their effectiveness in dealing with user queries that contain various types of ambiguity remains unknown, ultimately risking user trust and satisfaction. To this end, we introduce CLAMBER, a benchmark for evaluating LLMs using a well-organized taxonomy. Building upon the taxonomy, we construct 12K high-quality data to assess the strengths, weaknesses, and potential risks of various off-the-shelf LLMs.Our findings indicate the limited practical utility of current LLMs in identifying and clarifying ambiguous user queries, even enhanced by chain-of-thought (CoT) and few-shot prompting. These techniques may result in overconfidence …

Certified Policy Verification And Synthesis For Mdps Under Distributional Reach-Avoidance Properties, S. Akshay, Krishnendu Chatterjee, Tobias Meggendorfer, Dorde Zikelic

Research Collection School Of Computing and Information Systems

Markov Decision Processes (MDPs) are a classical model for decision making in the presence of uncertainty. Often they are viewed as state transformers with planning objectives defined with respect to paths over MDP states. An increasingly popular alternative is to view them as distribution transformers, giving rise to a sequence of probability distributions over MDP states. For instance, reachability and safety properties in modeling robot swarms or chemical reaction networks are naturally defined in terms of probability distributions over states. Verifying such distributional properties is known to be hard and often beyond the reach of classical state-based verification techniques. In …

Solving Long-Run Average Reward Robust Mdps Via Stochastic Games, Krishnendu Chatterjee, Ehsan Kafshdar Goharshady, Mehrdad Karrabi, Petr Novotný, Dorde Zikelic

Research Collection School Of Computing and Information Systems

Markov decision processes (MDPs) provide a standard framework for sequential decision making under uncertainty. However, MDPs do not take uncertainty in transition probabilities into account. Robust Markov decision processes (RMDPs) address this shortcoming of MDPs by assigning to each transition an uncertainty set rather than a single probability value. In this work, we consider polytopic RMDPs in which all uncertainty sets are polytopes and study the problem of solving long-run average reward polytopic RMDPs. We present a novel perspective on this problem and show that it can be reduced to solving long-run average reward turn-based stochastic games with finite state …

Peep With A Mirror: Breaking The Integrity Of Android App Sandboxing Via Unprivileged Cache Side Channel, Yan Lin, Joshua Wong, Xiang Li, Haoyu Ma, Debin Gao

Research Collection School Of Computing and Information Systems

Application sandboxing is a well-established security principle employed in the Android platform to safeguard sensitive information. However, hardware resources, specifically the CPU caches, are beyond the protection of this software-based mechanism, leaving room for potential side-channel attacks. Existing attacks against this particular weakness of app sandboxing mainly target shared components among apps, hence can only observe system-level program dynamics (such as UI tracing). In this work, we advance cache side-channel attacks by demonstrating the viability of non-intrusive and fine-grained probing across different app sandboxes, which have the potential to uncover app-specific and private program behaviors, thereby highlighting the importance of …

Epigenetic Dynamics In Zebrafish Development And Regeneration, Yujie Chen

Arts & Sciences Graduate Student Theses and Dissertations

Understanding the mechanisms that underlie cell fate and status decisions remains a complex task in biological research. Despite sharing identical genetic content, cells can differentiate into diverse types or remain in different statuses based on epigenetic dynamics. Here, I have structured my thesis into two distinct but interconnected parts, each focusing on a specific aspect of epigenetic control during zebrafish development and regeneration. The first part of the thesis focuses on the differentiation of pigment cells from neural crest cells to understand how the same progenitor cell population can bifurcate into two different pigment cell lineages. Through a comprehensive analysis …

Measuring Listening Effort In Bilingual Listeners For Speech Perception In Noise, Sita Carraturo

Arts & Sciences Graduate Student Theses and Dissertations

Noise is a common impedance to easy and accurate speech understanding. In the presence of noise, speech processing mechanisms proceed with partial or ambiguous inputs, and listeners will engage additional cognitive resources to make sense of what they hear. The extent to which this is situation is affected by diminished exposure to a language is an understudied question, but one that is relevant to bilingual listeners. By virtue of using two or more languages, bilinguals have less exposure to any one language than monolinguals of that language do. This, in turn, may make language processing more effortful for them, particularly …

Relationship Between School Counselor Multicultural Counseling Competence And School Counselor Consultation, Tamia H. Randolph-Alvarez

LSU New Orleans Theses and Dissertations

Professional counseling, as represented by associations responsible for setting standards and best practices in the profession, indicates that school counselors need to be prepared to engage in practices that result in systems-wide change. According to the Association for Counselor Education and Supervision [ACES] (2023), these practices include systems-based practices, multicultural competence, and consultation. However, research reveals that school counselors are not being prepared to utilize multicultural consultation (Chowela et al., 2020), defined as a systems-based practice (Hoffman et al., 2006) in this study. The Council for the Accreditation of Counseling and Related Educational Programs (2023) suggests that counseling graduate programs’ …

Experimental Investigations Of Elemental And Isotopic Partitioning With Applications To Geochronology, Lunar Crust, And Meteorite Erg Chech 002, Yuanyuan Liang

Arts & Sciences Graduate Student Theses and Dissertations

Elemental partitioning and crystallization processes on the Earth, Moon, and the meteorite Erg Chech 002 parent body are explored in this thesis, highlighting different mineral assemblage crystallization histories in three distinct geologic environments: the Earth's crust, the Moon's mantle, and the Erg Chech 002 (EC 002) parent body. Multiple series of one-atmosphere zircon synthesizing experiments, using basaltic andesite, andesite, and rhyolite starting materials, have been conducted (Chapter 2) to study the U/Th partitioning behavior between zircon and melt. The results reveal that sector zoning, fractional crystallization, typical igneous oxygen fugacities, and melt composition, particularly the whole rock silica content and …

Balancing Innovation And Integrity: Addressing Concerns Of Ai In Classroom Environments, Alexandria Patricia Thomas

Masters Projects

Research into the integration of artificial intelligence (AI) in education has highlighted critical issues such as bias, privacy, and plagiarism. Despite these challenges, many educational practices have not evolved to address them effectively. This project develops a comprehensive approach to preparing educators for ethical AI use, offering a series of lesson plans and resources (Appendices A-M) designed to enhance teachers' understanding and management of AI in the classroom. By providing practical tools and guidelines, the project supports educators in balancing technical skills with ethical considerations. It employs Paulo Freire’s Critical Pedagogy and the TPACK framework to address AI’s impact on …

A Retrospective Comparison Between Analytical Anisotropic Algorithm (Aaa) And Acuros Xb (Axb) Calculation Algorithms Concerning The Left Anterior Descending Coronary Artery Radiation Dose In Left Breast Cancer Patients Utilizing Deep Inspiration Breath Hold, Melissa Wilkinson

Masters Projects

Abstract

Background: Minimizing radiation exposure to the left anterior descending coronary artery (LAD) is crucial during left-sided breast cancer radiation therapy to prevent cardiac complications. Dose calculation algorithms can significantly impact the estimated LAD dose. This study investigates the differences between the Anisotropic Analytical Algorithm (AAA) and Acuros XB (AXB) algorithms in LAD dose calculations.

Methods: Thirty patients treated with radiation therapy after breast-conserving surgery were included. Dosimetric data for LAD, including mean and maximum dose, was calculated using AAA and AXB algorithms. The results were compared using statistical analysis.

Results: Statistically significant reductions in mean LAD dose were observed …

Increasing Secondary Students’ Spanish Oral Language Proficiency, Courtney Kozak

Masters Projects

Research has demonstrated the importance of students becoming orally proficient after taking foreign language classes. This project explores what can be done to combat the issue of students failing to gain an oral proficiency. It addresses instructional design and teacher impact, as well as student contribution to the issue. Multiple studies were reviewed on the best research backed practices for proficiency growth. After reading through all the research, it was determined that a proficiency-based instructional style that followed ACTFL World Readiness standards is the ideal approach. Going hand in hand with a proficiency-based instructional style is a classroom design of …

Girls First Fund: External Evaluation, Brian Medina Carranza, Dela Nai, Neelanjana Pandey, Grace Saul, Angel Del Valle, Fatima Zahra

Adolescents and Young People

The Population Council is leading an evaluation of programs funded by the Girls First Fund (GFF) in the Dominican Republic (DR), India, and Niger. These three countries were chosen to maximize geographic, social, cultural, and economic diversity across contexts, but common to all three is the continued persistence of child marriage and the presence of collective efforts to eliminate it. The broader goal of this evaluation is: 1) to demonstrate proof of concept for the impact of GFF-supported programs; 2) to examine key outcomes of interest to the GFF and the community-based organizations (CBOs) that it funds, including education, child …

High-Dose Intravenous Vitamin C Combined With Docetaxel In Men With Metastatic Castration-Resistant Prostate Cancer: A Randomized Placebo-Controlled Phase Ii Trial, Channing Paller, Marianna Zahurak, Adel Mandl, Nicole Metri, Aliya Lalji, Elisabeth Heath, William Kelly, Christopher Hoimes, Pedro Barata, Jason Taksey, Dominique Garrison, Kartick Patra, Ginger Milne, Nicole Anders, Julie Nauroth, Jennifer Durham, Catherine Marshall, Mark Markowski, Mario Eisenberger, Emmanuel Antonarakis, Michael Carducci, Samuel Denmeade, Mark Levine

Department of Medicine Faculty Papers

UNLABELLED: High-dose intravenous vitamin C (HDIVC) administered to produce pharmacologic concentrations shows promise in preclinical models and small clinical trials, but larger prospective randomized trials are lacking. We evaluated the clinical benefit of combining HDIVC with docetaxel in patients with progressive metastatic castration-resistant prostate cancer (mCRPC). In this double-blind, placebo-controlled phase II trial, 47 patients were randomized 2:1 to receive docetaxel (75 mg/m2 i.v.) with either HDIVC (1 g/kg) or placebo. Coprimary endpoints were PSA50 response and adverse event rates. Secondary endpoints included overall survival, radiographic progression-free survival, and quality of life measured using the Functional Assessment of Cancer Therapy-Prostate …

Examining The Relationship Between Recidivism And Community Size For Previously Incarcerated Individuals, Kayla A. Harbison

All-Inclusive List of Electronic Theses and Dissertations

Within urban contexts, there has been progress in understanding recidivism through the lens of both individual criminogenic risk factors and ecological factors. However, little research has examined how these factors apply across different community sizes. The current study attempted to address this gap by examining how recidivism differs across different-sized communities at the county level, measured both dichotomously (metro, nonmetro) and across a rural-urban continuum. This study also explored the relationship between specific types of offenses and recidivism across different-sized communities. Archival data from Indiana Department of Correction (IDOC) including 13,875 individuals released from prison within a five-year period was …

Changes In Revenues Associated With Antimicrobial Reimbursement Reforms In Germany, Matt Mcenany, Kevin Outterson

Faculty Scholarship

Policy declarations from the G7 and other high-level meetings call for increased incentives for antimicrobial research and development (R&D). Governments fund push incentives like CARB-X and GARDP, and G7 countries are now designing pull incentives—financial rewards given to manufacturers post-market authorization that are intended to encourage the creation and introduction of novel antimicrobials. Germany has declared previously at the G7 that it has developed a pull incentive that will increase revenues from sales of important new antimicrobials, principally by exempting them from some aspects of health technology benefit assessments and reference pricing, which should result in higher prices. This policy …

Evolutionary Genomic Analyses Of Canine E. Coli Infections Identify A Relic Capsular Locus Associated With Resistance To Multiple Classes Of Antimicrobials, Kristina Ceres, Jordan D. Zehr, Chloe Murrell, Jean K. Millet, Qi Sun, Holly C. Mcqueary, Alanna Horton, Casey Cazer, Kelly Sams, Guillaume Reboul, William B. Andreopoulos, Patrick K. Mitchell, Renee Anderson, Rebecca Franklin-Guild, Brittany D. Cronk, Bryce J. Stanhope, Claire R. Burbick, Rebecca Wolking, Laura Peak

Faculty Research, Scholarly, and Creative Activity

Infections caused by antimicrobial-resistant Escherichia coli are the leading cause of death attributed to antimicrobial resistance (AMR) worldwide, and the known AMR mechanisms involve a range of functional proteins. Here, we employed a pangenome wide association study (GWAS) approach on over 1,000 E. coli isolates from sick dogs collected across the US and Canada and identified a strong statistical association (empirical P < 0.01) of AMR, involving a range of antibiotics to a group 1 capsular (CPS) gene cluster. This cluster included genes under relaxed selection pressure, had several loci missing, and had pseudogenes for other key loci. Furthermore, this cluster is widespread in E. coli and Klebsiella clinical isolates across multiple host species. Earlier studies demonstrated that the octameric CPS polysaccharide export protein Wza can transmit macrolide antibiotics into the E. coli periplasm. We suggest that the CPS in question, and its highly divergent Wza, functions as an antibiotic trap, preventing antimicrobial penetration. We also highlight the high diversity of lineages circulating in dogs across all regions studied, the overlap with human lineages, and regional prevalence of resistance to multiple antimicrobial classes.

Energetics And Passivation Of Surface Defects In Hgte Colloidal Quantum Dots, Jacob D. Eisensmith

Theses

Colloidal quantum dots (CQDs), and especially HgTe CQDs, have the potential to disrupt traditional solid-state manufacturing of sensitive infrared photodetectors. Here we produce a surface model for (111) HgTe surfaces present in tetrahedral CQDs and attempt to use density functional theory (DFT) to explore various issues in the fabrication and performance of HgTe CQD photodetectors. One such concern is the presence of mid-gap trap states that can hurt device performance. In this work, we demonstrate that an abundant source of these trap states is unpassivated mercury at the surface of the nanocrystal, which can be corrected with passivation procedures. Furthermore, …

Development And Testing Of A Far Infrared Fourier Transform Spectrometer For The Calibration Of On-Chip Spectrometers, Dale Mercado

Theses

Astrophysical measurements in the far-infrared (FIR) wavelengths are crucial for a variety of astrophysical applications. These include probing the fuel for cosmic star formation, understanding the developmental cycles of protoplanetary disks, and investigating the role and composition of dust in the interstellar medium. Our collaboration is currently developing a FIR on-chip spectrometer that is operable in the range of 100-200µm, utilizing low-noise kinetic inductance detectors (KIDs). As these devices incorporate new technology, their spectral response must be measured and compared with traditional FIR optical dispersion methods. To this end, we have designed a cryogenic Fourier Transform Spectrometer (FTS) that operates …

Kilonova Modeling And Parameter Inference, Marko Ristic

Theses

This dissertation details the generation of a kilonova simulation training library and the training of subsequent machine learning emulators used for multidimensional interpolation. We begin with a Gaussian process interpolation methodology and apply that emulator for Bayesian inference with the AT 2017gfo light-curve dataset. We then expand on our emulator interpolation methodology, replacing the scalar Gaussian process emulator with a vector neural network emulator for light-curve generation. Our neural network emulator results are localized to a much more narrow region of the parameter space, prompting a detailed supportive investigation of the systematic uncertainty associated with our inference approach. We then …

Psychophysiological Insights Into Child-Centered Play Therapy For Trauma: A Case Study, Kristi L. Perryman, Samantha Robinson, Timothy J. Schoonover, Julia Conroy

Counseling, Leadership, and Research Methods Faculty Publications and Presentations

Existing literature thoroughly reviews the detrimental consequences that adverse childhood experiences (ACEs) have physically, emotionally, neurobiologically, and financially. It is imperative to develop effective treatments that offer a sense of hope to children who have been impacted. The established relationship between high ACE scores and physiological hyperarousal due to emotional dysregulation is clear in the literature. This relationship indicates that taking psychophysiological measures may be an effective method of gauging the effectiveness of trauma treatments. This study measured the heart rate of a child who had experienced multiple ACEs, during 16 child-centered play therapy (CCPT) sessions, using the Empatica (E4) …

Predicting Hurricane Evacuation Behavior Synthesizing Data From Travel Surveys And Social Media, Tanmoy Bhowmik, Naveen Eluru, Samiul Hasan, Aron Culotta, Kamol Chandra Roy

Civil and Environmental Engineering Faculty Publications and Presentations

Evacuation behavior models estimated using post-disaster surveys are not adequate to predict real-time dynamic population response as a hurricane unfolds. With the emergence of ubiquitous technology and devices in recent times, social media data with its higher spatio-temporal coverage has become a potential alternative for understanding evacuation behaviour during hurricanes. However, these data are often associated with selection bias and population representativeness issues. To that extent, the current study proposes a novel data fusion algorithm to combine heterogeneous data sources from transportation systems and social media, in a unified framework to understand and predict real-time population response during hurricanes. Specifically, …

Simulation Study On Confidence Interval Estimation For Standard Deviation With Non-Normal Distributions, Theophilus Oppong Kyeremeh

Electronic Theses and Dissertations

This study explores innovative approaches to constructing confidence intervals for the population standard deviation, σ, in non-normal data scenarios. While the sample standard deviation, s, is widely used, its reliability is compromised when dealing with skewed or heavy-tailed distributions and exhibits sensitivity to outliers. Our research addresses these limitations by investigating alternative estimation methods that offer greater robustness and accuracy.

Bowling With Bumper Rails: How Firearms Examiners Have Duped The Courts And Generated Low Error Rates Only By Avoiding Challenging Comparisons, Richard E. Gutierrez

UC Law Journal

No abstract provided.

From Uc Hastings To Uc Law Sf: An Examination Of The Renaming Process And Analysis Of Institutional Identity, Oliver Cheng

UC Law Journal

No abstract provided.