Physical Sciences and Mathematics Commons^™

Self-Optimizing Feature Generation Via Categorical Hashing Representation And Hierarchical Reinforcement Crossing, Wangyang Ying, Dongjie Wang, Kunpeng Liu, Leilei Sun, Yanjie Fu

Computer Science Faculty Publications and Presentations

Feature generation aims to generate new and meaningful features to create a discriminative representation space. A generated feature is meaningful when the generated feature is from a feature pair with inherent feature interaction. In the real world, experienced data scientists can identify potentially useful feature-feature interactions, and generate meaningful dimensions from an exponentially large search space in an optimal crossing form over an optimal generation path. But, machines have limited human-like abilities. We generalize such learning tasks as self-optimizing feature generation. Self-optimizing feature generation imposes several under-addressed challenges on existing systems: meaningful, robust, and efficient generation. To tackle these challenges, …

Conditional Optimal Sets And The Quantization Coefficients For Some Uniform Distributions, Evans Nyanney, Megha Pandey, Mrinal Kanti Roychowdhury

School of Mathematical and Statistical Sciences Faculty Publications and Presentations

Bucklew and Wise (1982) showed that the quantization dimension of an absolutely continuous probability measure on a given Euclidean space is constant and equals the Euclidean dimension of the space, and the quantization coefficient exists as a finite positive number. By giving different examples, in this paper, we have shown that the quantization coefficients for absolutely continuous probability measures defined on the same Euclidean space can be different. We have taken uniform distribution as a prototype of an absolutely continuous probability measure. In addition, we have also calculated the conditional optimal sets of n-points and the nth conditional quantization errors …

Hgprompt: Bridging Homogeneous And Heterogeneous Graphs For Few-Shot Prompt Learning, Xingtong Yu, Yuan Fang, Zemin Liu, Xinming Zhang

Research Collection School Of Computing and Information Systems

Graph neural networks (GNNs) and heterogeneous graph neural networks (HGNNs) are prominent techniques for homogeneous and heterogeneous graph representation learning, yet their performance in an end-to-end supervised framework greatly depends on the availability of task-specific supervision. To reduce the labeling cost, pre-training on selfsupervised pretext tasks has become a popular paradigm, but there is often a gap between the pre-trained model and downstream tasks, stemming from the divergence in their objectives. To bridge the gap, prompt learning has risen as a promising direction especially in few-shot settings, without the need to fully fine-tune the pre-trained model. While there has been …

Catnet: Cross-Modal Fusion For Audio-Visual Speech Recognition, Xingmei Wang, Jianchen Mi, Boquan Li, Yixu Zhao, Jiaxiang Meng

Research Collection School Of Computing and Information Systems

Automatic speech recognition (ASR) is a typical pattern recognition technology that converts human speeches into texts. With the aid of advanced deep learning models, the performance of speech recognition is significantly improved. Especially, the emerging Audio–Visual Speech Recognition (AVSR) methods achieve satisfactory performance by combining audio-modal and visual-modal information. However, various complex environments, especially noises, limit the effectiveness of existing methods. In response to the noisy problem, in this paper, we propose a novel cross-modal audio–visual speech recognition model, named CATNet. First, we devise a cross-modal bidirectional fusion model to analyze the close relationship between audio and visual modalities. Second, …

When Evolutionary Computation Meets Privacy, Bowen Zhao, Wei-Neng Chen, Xiaoguo Li, Ximeng Liu, Qingqi Pei, Jun Zhang

Research Collection School Of Computing and Information Systems

Recently, evolutionary computation (EC) has experienced significant advancements due to the integration of machine learning, distributed computing, and big data technologies. These developments have led to new research avenues in EC, such as distributed EC and surrogate-assisted EC. While these advancements have greatly enhanced the performance and applicability of EC, they have also raised concerns regarding privacy leakages, specifically the disclosure of optimal results and surrogate models. Consequently, the combination of evolutionary computation and privacy protection becomes an increasing necessity. However, a comprehensive exploration of privacy concerns in evolutionary computation is currently lacking, particularly in terms of identifying the object, …

Public Acceptance Of Using Artificial Intelligence-Assisted Weight Management Apps In High-Income Southeast Asian Adults With Overweight And Obesity: A Cross-Sectional Study, Han Shi Jocelyn Chew, Palakorn Achananuparp, Palakorn Achananuparp, Nicholas W. S. Chew, Yip Han Chin, Yujia Gao, Bok Yan Jimmy So, Asim Shabbir, Ee-Peng Lim, Kee Yuan Ngiam

Research Collection School Of Computing and Information Systems

Introduction: With in increase in interest to incorporate artificial intelligence (AI) into weight management programs, we aimed to examine user perceptions of AI-based mobile apps for weight management in adults with overweight and obesity. Methods: 280 participants were recruited between May and November 2022. Participants completed a questionnaire on sociodemographic profiles, Unified Theory of Acceptance and Use of Technology 2 (UTAUT2), and Self-Regulation of Eating Behavior Questionnaire. Structural equation modeling was performed using R. Model fit was tested using maximum-likelihood generalized unweighted least squares. Associations between influencing factors were analyzed using correlation and linear regression. Results: 271 participant responses were …

Vibmilk: Non-Intrusive Milk Spoilage Detection Via Smartphone Vibration, Yuezhong Wu, Wei Song, Yanxiang Wang, Dong Ma, Weitao Xu, Mahbub Hassan, Wen Hu

Research Collection School Of Computing and Information Systems

Quantifying the chemical process of milk spoilage is challenging due to the need for bulky, expensive equipment that is not user-friendly for milk producers or customers. This lack of a convenient and accurate milk spoilage detection system can cause two significant issues. First, people who consume spoiled milk may experience serious health problems. Secondly, milk manufacturers typically provide a “best before” date to indicate freshness, but this date only shows the highest quality of the milk, not the last day it can be safely consumed, leading to significant milk waste. A practical and efficient solution to this problem is proposed …

Transition-Informed Reinforcement Learning For Large-Scale Stackelberg Mean-Field Games., Pengdeng Li, Runsheng Yu, Xinrun Wang, Bo An

Research Collection School Of Computing and Information Systems

Many real-world scenarios including fleet management and Ad auctions can be modeled as Stackelberg mean-field games (SMFGs) where a leader aims to incentivize a large number of homogeneous self-interested followers to maximize her utility. Existing works focus on cases with a small number of heterogeneous followers, e.g., 5-10, and suffer from scalability issue when the number of followers increases. There are three major challenges in solving large-scale SMFGs: i) classical methods based on solving differential equations fail as they require exact dynamics parameters, ii) learning by interacting with environment is data-inefficient, and iii) complex interaction between the leader and followers …

Earnhft: Efficient Hierarchical Reinforcement Learning For High Frequency Trading, Molei Qin, Shuo Sun, Wentao Zhang, Haochong Xia, Xinrun Wang, Bo An

Research Collection School Of Computing and Information Systems

High-frequency trading (HFT) is using computer algorithms to make trading decisions in short time scales (e.g., second-level), which is widely used in the Cryptocurrency (Crypto) market, (e.g., Bitcoin). Reinforcement learning (RL) in financial research has shown stellar performance on many quantitative trading tasks. However, most methods focus on low-frequency trading, e.g., day-level, which cannot be directly applied to HFT because of two challenges. First, RL for HFT involves dealing with extremely long trajectories (e.g., 2.4 million steps per month), which is hard to optimize and evaluate. Second, the dramatic price fluctuations and market trend changes of Crypto make existing algorithms …

M3sa: Multimodal Sentiment Analysis Based On Multi-Scale Feature Extraction And Multi-Task Learning, Changkai Lin, Hongju Cheng, Qiang Rao, Yang Yang

Research Collection School Of Computing and Information Systems

Sentiment analysis plays an indispensable part in human-computer interaction. Multimodal sentiment analysis can overcome the shortcomings of unimodal sentiment analysis by fusing multimodal data. However, how to extracte improved feature representations and how to execute effective modality fusion are two crucial problems in multimodal sentiment analysis. Traditional work uses simple sub-models for feature extraction, and they ignore features of different scales and fuse different modalities of data equally, making it easier to incorporate extraneous information and affect analysis accuracy. In this paper, we propose a Multimodal Sentiment Analysis model based on Multi-scale feature extraction and Multi-task learning (M 3 SA). …

Leveraging Llms And Generative Models For Interactive Known-Item Video Search, Zhixin Ma, Jiaxin Wu, Chong-Wah Ngo

Research Collection School Of Computing and Information Systems

While embedding techniques such as CLIP have considerably boosted search performance, user strategies in interactive video search still largely operate on a trial-and-error basis. Users are often required to manually adjust their queries and carefully inspect the search results, which greatly rely on the users’ capability and proficiency. Recent advancements in large language models (LLMs) and generative models offer promising avenues for enhancing interactivity in video retrieval and reducing the personal bias in query interpretation, particularly in the known-item search. Specifically, LLMs can expand and diversify the semantics of the queries while avoiding grammar mistakes or the language barrier. In …

Frameworks For Measuring Population Health: A Scoping Review, Sze Ling Chan, Clement Zhong Hao Ho, Nang Ei Ei Khaing, Ezra Ho, Candelyn Pong, Jia Sheng Guan, Calida Chua, Zongbin Li, Trudi Lim Wenqi, Sean Shao Wei Lam, Lian Leng Low, Choon How How

Research Collection School Of Computing and Information Systems

Introduction Many regions in the world are using the population health approach and require a means to measure the health of their population of interest. Population health frameworks provide a theoretical grounding for conceptualization of population health and therefore a logical basis for selection of indicators. The aim of this scoping review was to provide an overview and summary of the characteristics of existing population health frameworks that have been used to conceptualize the measurement of population health. Methods We used the Population, Concept and Context (PCC) framework to define eligibility criteria of frameworks. We were interested in frameworks applicable …

Delving Into Multimodal Prompting For Fine-Grained Visual Classification, Xin Jiang, Hao Tang, Junyao Gao, Xiaoyu Du, Shengfeng He, Zechao Li

Research Collection School Of Computing and Information Systems

Fine-grained visual classification (FGVC) involves categorizing fine subdivisions within a broader category, which poses challenges due to subtle inter-class discrepancies and large intra-class variations. However, prevailing approaches primarily focus on uni-modal visual concepts. Recent advancements in pre-trained vision-language models have demonstrated remarkable performance in various high-level vision tasks, yet the applicability of such models to FGVC tasks remains uncertain. In this paper, we aim to fully exploit the capabilities of cross-modal description to tackle FGVC tasks and propose a novel multimodal prompting solution, denoted as MP-FGVC, based on the contrastive language-image pertaining (CLIP) model. Our MP-FGVC comprises a multimodal prompts …

Artificial Intelligence Model Predicts Sudden Cardiac Arrest Manifesting With Pulseless Electric Activity Versus Ventricular Fibrillation, Lauri Holmstrom, Bryan Bednarski, Harpriya Chugh, Habiba Aziz, Hoang Nhat Pham, Arayik Sargsyan, Audrey Uy-Evanado, Damini Dey, Angelo Salvucci, Jonathan Jui, Kyndaron Reinier, Piotr J Slomka, Sumeet S Chugh

Student and Faculty Publications

BACKGROUND: There is no specific treatment for sudden cardiac arrest (SCA) manifesting as pulseless electric activity (PEA) and survival rates are low; unlike ventricular fibrillation (VF), which is treatable by defibrillation. Development of novel treatments requires fundamental clinical studies, but access to the true initial rhythm has been a limiting factor.

METHODS: Using demographics and detailed clinical variables, we trained and tested an AI model (extreme gradient boosting) to differentiate PEA-SCA versus VF-SCA in a novel setting that provided the true initial rhythm. A subgroup of SCAs are witnessed by emergency medical services personnel, and because the response time is …

Examination Of Traditional Botnet Detection On Iot-Based Bots, Ashley Woodiss-Field, Michael N. Johnstone, Paul Haskell-Dowland

Research outputs 2022 to 2026

A botnet is a collection of Internet-connected computers that have been suborned and are controlled externally for malicious purposes. Concomitant with the growth of the Internet of Things (IoT), botnets have been expanding to use IoT devices as their attack vectors. IoT devices utilise specific protocols and network topologies distinct from conventional computers that may render detection techniques ineffective on compromised IoT devices. This paper describes experiments involving the acquisition of several traditional botnet detection techniques, BotMiner, BotProbe, and BotHunter, to evaluate their capabilities when applied to IoT-based botnets. Multiple simulation environments, using internally developed network traffic generation software, were …

Analysis Of The Outer Retinal Bands In Abca4 And Prph2-Associated Retinopathy Using Oct, Rachael C. Heath Jeffery, Johnny Lo, Jennifer A. Thompson, Tina M. Lamey, Terri L. Mclaren, John N. De Roach, Lauren N. Ayton, Andrea L. Vincent, Abhishek Sharma, Fred K. Chen

Research outputs 2022 to 2026

Purpose: To evaluate the outer retinal bands using OCT in ABCA4- and PRPH2-associated retinopathy and develop a novel imaging biomarker to differentiate between these 2 genotypes. Design: Multicenter case-control study. Participants: Patients with a clinical and genetic diagnosis of ABCA4- or PRPH2-associated retinopathy and an age-matched control group. Methods: Macular OCT was used to measure the thickness of the outer retinal bands 2 and 4 by 2 independent examiners at 4 retinal loci. Main Outcome Measures: Outcome measures included the thicknesses of band 2, band 4, and the band 2/band 4 ratio. Linear mixed modeling was used to make comparisons …

Machine Learning-Enhanced All-Photovoltaic Blended Systems For Energy-Efficient Sustainable Buildings, Mohammad Nur-E-Alam, Kazi Z. Mostofa, Boon K. Yap, Mohammad K. Basher, Mohammad A. Islam, Mikhail Vasiliev, Manzoore E. M. Soudagar, Narottam Das, Tiong S. Kiong

Research outputs 2022 to 2026

The focus of this work is on the optimization of an all-photovoltaic hybrid power generation systems for energy-efficient and sustainable buildings, aiming for net-zero emissions. This research proposes a hybrid approach combining conventional solar panels with advanced solar window systems and building integrated photovoltaic (BIPV) systems. By analyzing the meteorological data and using the simulation models, we predict energy outputs for different cities such as Kuala Lumpur, Sydney, Toronto, Auckland, Cape Town, Riyadh, and Kuwait City. Although there are long payback times, our simulations demonstrate that the proposed all-PV blended system can meet the energy needs of modern buildings (up …

Edsuch: A Robust Ensemble Data Summarization Method For Effective Medical Diagnosis, Mohiuddin Ahmed, A. N. M. B. Rashid

Research outputs 2022 to 2026

Identifying rare patterns for medical diagnosis is a challenging task due to heterogeneity and the volume of data. Data summarization can create a concise version of the original data that can be used for effective diagnosis. In this paper, we propose an ensemble summarization method that combines clustering and sampling to create a summary of the original data to ensure the inclusion of rare patterns. To the best of our knowledge, there has been no such technique available to augment the performance of anomaly detection techniques and simultaneously increase the efficiency of medical diagnosis. The performance of popular anomaly detection …

Local Habitat Composition And Complexity Outweigh Seascape Effects On Fish Distributions Across A Tropical Seascape, Molly Moustaka, Richard D. Evans, Gary A. Kendrick, Glenn A. Hyndes, Michael V. W. Cuttler, Tahlia J. Bassett, Michael J. O’Leary, Shaun K. Wilson

Research outputs 2022 to 2026

Context: The distribution of animals is influenced by a complex interplay of landscape, environmental, habitat, and anthropogenic factors. While the effects of each of these forces on fish assemblages have been studied in isolation, the implications of their combined influence within a seascape remain equivocal. Objectives: We assessed the importance of local habitat composition, seascape configuration, and environmental conditions for determining the abundance, diversity, and functional composition of fish assemblages across a tropical seascape. Methods: We quantified fish abundance in coral, macroalgal, mangrove, and sand habitats throughout the Dampier Archipelago, Western Australia. A full-subsets modelling approach was used that incorporated …

Evaluating The Reliability And Accuracy Of Alchemical Binding Free Energy Methods And Calculations, Fnu Sheenam

Dissertations, Theses, and Capstone Projects

Molecular recognition plays a crucial role in various biological processes, such as enzymatic reactions, signal transduction, and genetic information processing. Investigating how proteins selectively bind to their partners is an active research area, but there is a lack of experimental details on protein structures and interactions in molecular complexes. Computational techniques based on macromolecular structures offer a way to predict protein-ligand interactions and explore their recognition mechanisms. Estimating binding affinities, particularly through alchemical binding free energy calculations, has become valuable in supporting drug discovery. This work introduces new methodologies, utilizing the Alchemical Transfer Method, to address issues like poor convergence …

Examining The Health Risks Of Particulate Matter 2.5 In New York City: How It Affects Marginalized Groups And The Steps Needed To Reduce Air Pollution, Freddy Castro

Dissertations, Theses, and Capstone Projects

The following examines the impact of particulate matter 2.5 (PM2.5) on public health, focusing on its sources and effects on vulnerable populations in New York City. PM2.5 is a particle that is 2.5 micrometers or less in diameter and, because of its size, can enter the bloodstream affecting the respiratory and cardiovascular systems and further complicating the health of the immunocompromised. Recent studies have shown that PM2.5 can come from various sources, including transportation and industrial emissions, as well as indoor sources like cigarettes and gas-operated stoves. Despite reduced levels of PM2.5 due to recent policy changes and initiatives taken …

Low Shear In Short-Term Impacts Endothelial Cell Traction And Alignment In Long-Term, Mohanish K. Chandurkar, Nikhil Mittal, Shaina P. Royer-Weeden, Steven D. Lehmann, Yeonwoo Rho, Sangyoon J. Han

Michigan Tech Publications, Part 2

Within the vascular system, endothelial cells (ECs) are exposed to fluid shear stress (FSS), a mechanical force exerted by blood flow that is critical for regulating cellular tension and maintaining vascular homeostasis. The way ECs react to FSS varies significantly; while high, laminar FSS supports vasodilation and suppresses inflammation, low or disturbed FSS can lead to endothelial dysfunction and increase the risk of cardiovascular diseases. Yet, the adaptation of ECs to dynamically varying FSS remains poorly understood. This study focuses on the dynamic responses of ECs to brief periods of low FSS, examining its impact on endothelial traction—a measure of …

Stereospecific Cross-Coupling Reactions Of Vinyl Triflates And An Investigation Of The Synthesis And Reactivity Of Activated Alkyltricyclohexylstannanes, Pejman Ghaemimohammadi

Dissertations, Theses, and Capstone Projects

Organic chemistry exists in all aspects of everyday life from polymers to pharmaceutics. Formation of carbon-carbon bonds is essential for the synthesis of complex organic compounds. This goal can be achieved by using transition metal catalysts. Second-row transition metals such as Pd, Pt, Rh, and Ir have shown remarkable efficacy in these reactions. Metal-catalyzed cross-coupling reactions form carbon-carbon bonds between electrophiles and nucleophiles using transition metal catalysts. Pd is mainly used in this reaction.

Transition metal-catalyzed cross-coupling reactions that form a bond between two C(sp²) carbons have been widely studied over the past decades. C(sp²)-C(sp^{3 …}

Low Elevation Coastal Zone (Lecz) Population Social Vulnerability And Risk: A Spatial Analysis Based On The 2015-16 National Family Health Survey (Nfhs-4) Of India, Paradorn Wongchanapai

Dissertations, Theses, and Capstone Projects

Background & Problem Statement:

This study assessed social vulnerability and risk to hydroclimate hazards of Indian urban and rural populations in the Low Elevation Coastal Zone (LECZ), a contiguous coastal area with elevations less than 10 meters. The LECZ is considered an at-risk area as hydroclimate hazards tend to be heightened. Moreover, being the most populous country in the world, any hydroclimate hazards that hit the LECZ would disproportionately impact a large number of Indian populations. Understanding the social vulnerability and risk of the LECZ populations can help mitigate and prevent potential adverse outcomes for the populations.

Importantly, this study …

Clustering Of Patients With Heart Disease, Mukadder Cinar

Dissertations, Theses, and Capstone Projects

Heart disease, a leading cause of mortality worldwide, presents complex challenges in public health due to its varied manifestations. Accurate diagnosis and patient stratification are essential for effective management and improved outcomes. In response, this study employed machine learning techniques to analyze heart disease data obtained from UCI Machine Learning Repository, aiming to enhance patient care through advanced data analysis.

The study began with the application of K-Nearest Neighbors (KNN) classification, which categorized patients into 'Disease' and 'No Disease' groups. This preliminary step provided initial insights into the structure of the dataset. Subsequently, K-means clustering was applied in two rounds, …

Making Sense Of Making Parole In New York, Alexandra Mcglinchy

Dissertations, Theses, and Capstone Projects

For many individuals incarcerated in New York, the initial step toward freedom begins with an interview with the Board of Parole. This process, however, is frequently a complex and challenging one, characterized by repeated denials and extended incarcerations. The disparity in outcomes – where one individual may receive over 20 denials and another is granted parole on their first attempt – highlights the ambiguity and inconsistency in the parole decision-making process. This project aims to clarify the factors that influence parole decisions by concentrating on measurable variables. These include age, race, duration of sentence served, proportion of sentence served, type …

Studies Of Oxidopyrylium Ylide Dimers And Their Utility In The Synthesis Of Densely Functionalized Troponoids, Daniel V. Schiavone

Dissertations, Theses, and Capstone Projects

Herein, we describe our efforts to investigate mechanistic considerations of oxidopyrylium ylides and their use in cycloadditions to form complex bicyclic structures and the advancement of these cycloadducts to the synthesis of densely functionalized tropolones. This work also describes applications of tropolones as biologically relevant structures and dynamic fluorophores.

In Chapter 1, we expand upon previous studies done in the Murelli Lab describing the ability of 3-hydroxy-4-pyrone derived oxidopyrylium dimers to revert back to ylides in situ and as result can be used as a clean oxidopyrylium ylide source in accessing oxabicyclic compounds. Intermolecular cycloadditions of these ylides can be …

Containerization Of Seafarers In The International Shipping Industry: Contemporary Seamanship, Maritime Social Infrastructures, And Mobility Politics Of Global Logistics, Liang Wu

Dissertations, Theses, and Capstone Projects

This dissertation discusses the mobility politics of container shipping and argues that technological development, political-economic order, and social infrastructure co-produce one another. Containerization, the use of standardized containers to carry cargo across modes of transportation that is said to have revolutionized and globalized international trade since the late 1950s, has served to expand and extend the power of international coalitions of states and corporations to control the movements of commodities (shipments) and labor (seafarers). The advent and development of containerization was driven by a sociotechnical imaginary and international social contract of seamless shipping and cargo flows. In practice, this liberal, …

Deep Learning-Based Human Action Understanding In Videos, Elahe Vahdani

Dissertations, Theses, and Capstone Projects

The understanding of human actions in videos holds immense potential for technological advancement and societal betterment. This thesis explores fundamental aspects of this field, including action recognition in trimmed clips and action localization in untrimmed videos. Trimmed videos contain only one action instance, with moments before or after the action excluded from the video. However, the majority of videos captured in unconstrained environments, often referred to as untrimmed videos, are naturally unsegmented. Untrimmed videos are typically lengthy and may encompass multiple action instances, along with the moments preceding or following each action, as well as transitions between actions. In the …

Rational Design Of Peptide-Based Materials Informed By Multiscale Molecular Dynamics Simulations, Dhwanit Rahul Dave

Dissertations, Theses, and Capstone Projects

The challenge of establishing a sustainable and circular economy for materials in medicine and technology necessitates bioinspired design. Nature's intricate machinery, forged through evolution, relies on a finite set of biomolecular building blocks with through-bond and through-space interactions. Repurposing these molecular building blocks requires a seamless integration of computational modeling, design, and experimental validation. The tools and concepts developed in this thesis pioneer new directions in peptide-materials design, grounded in fundamental principles of physical chemistry. We present a synergistic approach that integrates experimental designs and computational methods, specifically molecular dynamics simulations, to gain in-depth molecular insights crucial for advancing the …