Physical Sciences and Mathematics Commons^™

A Comparative Analysis Of Explainable Ai (Xai) Techniques For Transparent And Reliable Image Classification, Sovon Chakraborty, Shakib Mahmud Dipto, Kevin R. Pilkiewicz, Michael L. Mayo, Pratip Rana

Computer Science Faculty Publications

Evaluating the trustworthiness of black-box machine learning models remains a significant methodological challenge. Their lack of transparency and interpretability limits applicability, because stakeholders often seek transparency before trusting the results of black-box machine learning models. Explainable AI (XAI) methods provide for human-understandable justifications and informed decision-making of these black-box architectures. Therefore, it is imperative to select the proper XAI model tailored to specific tasks. In this research, we focus on examining four XAI techniques: PEEK, LRP, GRAD-CAM, and LIME to understand how they perform against each other for image classification tasks. We evaluate the performance, robustness, generalizability, noise stability, and …

Attf-Gnn: An Attention-Based Multi-Omics Graph Neural Network With Modality Learning For Disease Subtyping, Sovon Chakraborty, Eleni Adam, Terry Stilwell, Harold Riethman, Desh Ranjan, Pratip Rana

Computer Science Faculty Publications

We propose AttF-GNN, an attention-based graph fusion strategy for diseases classification and subtyping. In multiomics analysis, not all types of molecular data are equally relevant for disease subtyping and considering all modalities equally may obscure discriminative signals and limit the effectiveness of predictive models by overlooking modality-specific contributions. Therefore, we design an attention-based multimodal GraphSAGE framework that can automatically emphasize the modalities providing the most relevant information for classification. At first, we have constructed three graphs using mRNA, RNA-seq and DNA methylation modalities, and train each omics with individual GraphSAGE encoders. Next, a unified intersection graph is formed using an …

Attention-Based Multi-Omics Fusion For Drug Synergy Prediction, Kusal Debnath, Pratip Rana, Preetam Ghosh

Computer Science Faculty Publications

Drug combination therapy in disease management gained popularity in the last few decades. Computational modeling of such combinations is an active area of research in the drug discovery domain. While earlier approaches solely emphasized on the structural features of participating drugs for designing synergistic models, they lack other crucial factors directly linked with drug administration - omics expressions. As differential omics expression is a downstream consequence of the administered drug combinations, utilizing such expressions while designing synergistic models promises robust and dynamic modeling. In this work, we propose SynergyLM that fuses multi-omics features with drug embeddings to build an omics-aware …

Optimal Control Analysis Of Malaria Transmission In The Presence Of Insecticide Resistance And Climate Variability In Kenya, Lorna Chepkemoi, Titus Okello Orwa, Samuel Mwalili, Rachel Waema Mbogo, Steeven Belvinos Affognon, Daisy Salifu, Henri E.Z. Tonnang

All Peer-Reviewed Publications

Malaria remains a major public health concern in Kenya, where changing climatic conditions, insecticide resistance, and mosquito behavioral adaptations continue to challenge control efforts. This study develops and analyzes a climate-sensitive malaria transmission model that incorporates mosquito behavior, insecticide resistance, and vector–human ecological dynamics to identify optimal control strategies for the Kenyan context. The model’s well-posedness was established, the basic reproduction number (R0) computed using the next-generation matrix method, and the stability of equilibrium points assessed. Spatial analysis of R0 was performed using temperature and rainfall raster data to map transmission risk across Kenya under varying insecticide use scenarios. Results …

Early Icu Physiological Subtyping From Time-Series Data: A Comparative Study Of Feature Complexity, Predictive Performance, And Model Interpretability, Rojeena Khadka

College of Graduate Studies: Theses & Dissertations

Intensive Care Unit (ICU) patients do not follow a single uniform physiological pattern. Patients admitted with the same diagnosis show different clinical trajectories over time making standardized classification and treatment approaches insufficient. The increasing availability of large-scale electronic health records in MIMIC-IV makes it possible to investigate such heterogeneity through data-driven approach that captures how physiology evolves during the early phase of ICU admission. This thesis compares two analytical pipelines designed to identify physiological subtypes from the first 48 hours of ICU time series data. This study then assesses how well these subtypes predict in-hospital mortality. The first approach, referred …

Designing Narrative-Based Ai Assistance For Sensemaking In Collaborative Environments: Case Studies In Education And Dementia Care, Dylan Edward Moore

Dartmouth College Ph.D Dissertations

This thesis addresses a gap in the human-computer interaction literature regarding the design, development, and evaluation of narrative-based AI assistance for collaborative, complex problem solving. I explore this design space through three case studies across the domains of education and dementia care. This work encompasses multi-year industry partnerships and longitudinal fieldwork, user-centered design, dataset curation, model training, and system evaluation.

Specifically, the first case study considers a story-based web platform for teaching AI literacy through peer-generated, personalized narrative scaffolding. Learners on the platform showed significant knowledge gains and other learning-related outcomes. To describe the novel design of this system, I …

Self-Calibrating Uav Navigation: Reinforcement Learning Approaches For Horizontal Trajectory Estimation, Shirin Nasr-Esfahani, S. Jagannathan

Electrical and Computer Engineering Faculty Research & Creative Works

Accurate unmanned aerial vehicle (UAV) trajectory estimation is essential for autonomous navigation, particularly in GPS-denied environments. Visualodometry and simultaneous localization and mapping (SLAM) approaches require precise camera intrinsic parameters, which are typically obtained through predefined or offline calibration. Instead, in this work, we propose a reinforcement learning (RL)-based self-calibration framework that estimates camera intrinsic parameters directly from monocular video sequences, without requiring prior knowledge of the camera, environment, or calibration targets. This intrinsic parameter estimation is then leveraged to achieve robust UAV trajectory estimation using only video data. We formulate the problem as a sequential decision-making task, where an RL …

Mtl_Tx: A Multi-Task Transformer Model For Improved Radiation Time-Series Estimation, Hongfang Zhang, Adam Stavola, Hal Ferguson, Bence Budavari, Hongyi Wu, Chiman Kwan, Jiang Li

Electrical & Computer Engineering Faculty Publications

Controlling radiation doses at potential radioactive facilities is critical to ensuring the safety of both personnel and the public. At the Thomas Jefferson National Accelerator Facility (JLab), multiple sensors are deployed around the three experimental halls to monitor key parameters, including single-beam current, energy levels, current leakage, and radiation values during accelerator operations. In this study, we developed a Multi-task Transformer model, MTL_TX, to accurately estimate radiation doses at sensor locations based on historical data, with the aim of enhancing safety in accelerator facilities and surrounding public areas. To improve estimation accuracy, we integrated two innovative components into the proposed …

Multimodal Machine Learning For Alzheimer's Disease Classification Using Adni Biomarker Fusion, Hesameddin Mostaghimi, Hamid R. Okhravi, Bahar Niknejad, Daniel A. Cohen, Michel A. Audette, Alzheimer's Disease Neuroimaging Initiative

Electrical & Computer Engineering Faculty Publications

Despite ongoing advances, accurate diagnosis of Alzheimer’s disease (AD) remains challenging due to its multifactorial nature, comorbidities, and clinical heterogeneity. Accordingly, approaches that combine multimodal data may improve AD classification by integrating complementary information. To investigate this, we evaluated classification performance using a preprocessed ADNI-3 dataset comprising a shared set of clinical/cognitive features along with four imaging modality-based cohorts: trimodal (MRI + amyloid PET + tau PET), MRI + amyloid PET, MRI + tau PET, and MRI-only. We trained a range of supervised machine learning (ML) and deep learning (DL) classifiers using stratified five-fold cross-validation and evaluated performance using accuracy, …

Standardization Of Neuromuscular Reflex Analysis—Role Of Fine-Tuned Vision-Language Model Consortium And Openai Gpt-Oss Reasoning Llm-Enabled Decision Support System, Eranga Bandara, Ross Gore, Sachin Shetty, Ravi Mukkamala, Christopher K. Rhea, Brittany S. Samulski, Amin Hass, Atmaram Yarlagadda, Shaifali Kaushik, Malith De Silva, Andriy Maznychenko, Inna Sokolowska, Kasun De Zoysa

VMASC Publications

Background/Objectives: Accurate assessment of neuromuscular reflexes, such as the Hoffmann reflex (H-reflex), plays a critical role in sports science, rehabilitation, and clinical neurology. Conventional interpretation of H-reflex electromyography (EMG) waveforms is subject to inter-rater variability and interpretive bias, limiting reliability and standardization. This study aims to develop an automated, interpretable, and robust agentic AI–driven framework for H-reflex waveform analysis. Methods: We propose a fine-tuned Vision–Language Model (VLM) consortium combined with a reasoning Large Language Model (LLM)–enabled decision support system for automated H-reflex interpretation. Multiple VLMs were fine-tuned on curated datasets of H-reflex EMG waveform images annotated with expert clinical observations, …

Machine Learning Classification Of Prostate Cancer Genomic Sequences Using K-Mer And Sequence-Derived Features, Kuldeep Rawat, Hirendra Nath Banerjee, Jamie Noble, Saa Naudia Deloatch, Satyendra Banerjee, Sachin Shetty, Soumya Banerjee

VMASC Publications

Prostate cancer disproportionately impacts African American men, who experience significantly higher mortality rates and earlier disease onset than other populations. Current diagnostic approaches, including prostate-specific antigen testing and biopsy, lack sufficient specificity and sensitivity, underscoring the need for accurate, molecular-level classification tools. This paper presents a machine learning framework for binary classification of genomic DNA sequences as cancerous or healthy. A dataset of 1684 FASTA-formatted sequences obtained from the National Library of Medicine - GenBank was analyzed, with 1662 sequences retained after quality control filtering. Feature engineering yielded 67 attributes, including GC content, Shannon entropy, sequence length, and trinucleotide k-mer …

Managing Multi-Drug Resistance: An Evolutionary Game Theory And Optimal Control Approach, Shukhrat Nasrulloev

Theses and Dissertations

Multi-drug resistance is an evolutionary process in which treatment eliminates sensitive cells, allowing resistant clones to dominate. This thesis investigates this process using a framework integrating population dynamics, evolutionary game theory, and optimal control theory. We develop a two-population logistic growth model describing competition between drug-sensitive and drug-resistant cells under treatment, construct dose-dependent payoff matrices and replicator dynamics to characterize evolutionary competition, and derive a critical drug level D_crit = (r_S - rR)/(dS - d_R) at which resistant cells gain a fitness advantage. An optimal control problem is formulated via Pontryagin's Maximum Principle to identify schedules …

Lessons On Generative Artificial Intelligence From The American Association Of Dental Editors And Journalists (Aadej), Christopher J. Smiley

Journal of the American College of Dentists

The American Association of Dental Editors and Journalists (AADEJ) recently released "Guidance for Authors, Editors and Publishers on the Use of Generative AI". Developed by an 11-member stakeholder panel, this guidance paper serves a dual purpose: It provides practical strategies for mitigating the risks that generative artificial intelligence (GAI) poses to professional writing and explains how GAI creates these risks. Understanding both the underlying vulnerabilities of GAI and how to address them is critical for users at all levels, from authors and reviewers to editors, publishers, and general users, to maintain the validity and reliability of their written work. This …

The Role Of Resistance Training In Managing Menopause-Related Changes, Alyson K. Strauss

Honors Undergraduate Theses

Resistance training serves as a critical non-pharmacological strategy for mitigating the physiological and psychological declines associated with the menopausal transition. This natural biological phase, marked by declining estrogen levels, typically leads to accelerated bone mineral density loss, sarcopenia, metabolic regulation disturbance, and heightened mood disturbances. To evaluate the efficacy of targeted exercise interventions, a systematic review was conducted following PRISMA guidelines. From an initial pool of 60 relevant articles, 13 studies met the specific inclusion criteria, focusing on structured resistance training protocols for perimenopausal, menopausal, and postmenopausal cohorts.

The synthesis of these studies indicates that resistance training is a highly …

Safety Aware Continual Reinforcement Learning-Based Output Tracking Control Of Nonlinear Continuous-Time Systems, Irfan Ganie, Sarangapani Jagannathan

Electrical and Computer Engineering Faculty Research & Creative Works

An output feedback (OF)-based control scheme utilizing both a scalable multilayer neural network (MNN) observer and actor–critic MNN via integral reinforcement learning (IRL)/adaptive dynamics programming (ADP) approach for a class of nonlinear systems with output constraints is introduced. The proposed observer, critic, and actor MNN weight updates are derived using a singular value decomposition (SVD) of MNN activation function gradient along with output error, Bellman and control input errors, respectively. Next, the approach incorporates continual learning (CL), utilizing a penalty function in the weight update laws for both actor–critic MNNs to consolidate knowledge from previous tasks and enhance learning in …

Microgravity-Induced Alterations In Left Atrial Hemodynamics And Thrombogenic Risk: Insights From Healthy And Atrial Fibrillation Models, Grace M. Hoeppner

Dissertations, Master's Theses and Master's Reports

Background: Microgravity exposure alters cardiovascular loading, yet its impact on left atrial flow dynamics and thrombotic risk remains poorly understood. This study investigates how spaceflight-relevant microgravity-induced changes in cardiac outflow affect left atrial hemodynamics in healthy individuals and patients with atrial fibrillation.

Methods: Patient-specific left atrial models were generated for three healthy individuals and three AF patients. Computational fluid dynamics (CFD) simulations were performed using each patient’s baseline mitral outflow waveform and two modified waveforms representing short- and long-duration post-flight cardiac loading changes derived from echocardiographic observations. Hemodynamic metrics included left atrial velocity, time averaged wall shear stress, oscillatory shear …

Approximation Of Solution Of Certain Nonlinear Problems With Applications, Faisal Aijaz

Master of Science in Mathematics

In the thesis, we established the strong convergence of inertial iterative scheme designed to approximate the common solutions of the equilibrium and the fixed-point problem for non-expansive mappings in Hilbert spaces. Through extensive numerical analysis, we analyze the efficiency of the algorithm and demonstrate that the incorporation of the inertial technique yields a significantly faster convergence rate than existing comparable schemes. Furthermore, we examine the effect of initial value on the convergence dynamics, offering visual and comparative assessments against classical methods. To highlight the practical utility and versatility of our framework, we apply the algorithm to signal restoration problems, and …

Integrative Machine Learning Of Genetic And Lifestyle Factors For Personalized Skin Health, Yassine Benachour, Lina Maloukh, Barbara Geusens

All Works

Objective: To develop an AI framework that combines genetic, phenotypic, and lifestyle data for profiling skin-health patterns and generating hypothesis-supporting summaries for potential decision support. Methods and procedures: A dataset of 5,254 individuals integrates six genes (FLG, AQP3, MMP-1, MMP-3, SOD2, GPX), six phenotype severities, and 20+ lifestyle factors. Mutation burden and interactions are tested by ANOVA. K-modes clustering identifies four interpretable dermatological profiles within the cohort and is embedded in leakage-free nested cross-validation (train-only selection; test labels from training centroids). Subtypes are predicted from genetics plus lifestyle using an XGBoost (XGB) classifier; explainability uses gain, permutation importance, and SHAP …

Memoryart: Enhancing Llms Via Multi-Memory Models With Adaptive Resonance Theory For Healthcare Agents, Renke Dai, Hebin Hu, Jiahui Zhang, Yilin Kang, Ah-Hwee Tan

Research Collection School Of Computing and Information Systems

Though promising in healthcare consultation applications, large language models (LLMs) face critical limitations in retaining and utilizing long-term memory across multiturn interactions. In particular, existing memory enhancing paradigms are constrained by limited context windows and embedding-based retrieval, often failing to maintain task relevance and still suffering from memory prototype collapse in multi-turn healthcare consultation. To address these challenges, we propose a cognitively-inspired memory framework named MemoryART, which is grounded in Adaptive Resonance Theory (ART)—a cognitive and learning theory of how humans and animals adapt to dynamic environments. MemoryART employs three memory modules—working memory, episodic memory, and semantic memory to support …

Interpretable Multimodal Zero Shot Ecg Diagnosis Via Structured Clinical Knowledge Alignment, Jialu Tang, Hung Manh Pham, Ignace De Lathauwer, Henk S. Schipper, Yuan Lu, Dong Ma, Aaqib Saeed

Research Collection School Of Computing and Information Systems

Electrocardiogram (ECG) interpretation is essential for cardiovascular disease diagnosis, but current automated systems often struggle with transparency and generalization to unseen conditions. To address this, we introduce ZETA, a zero-shot multimodal framework designed for interpretable ECG diagnosis aligned with clinical workflows. ZETA uniquely compares ECG signals against structured positive and negative clinical observations, which are curated through an LLM-assisted, expertvalidated process, thereby mimicking differential diagnosis. Our approach leverages a pre-trained multimodal model to align ECG and text embeddings without disease-specific fine-tuning. Empirical evaluations demonstrate ZETA’s competitive zero-shot classification performance and, importantly, provide qualitative and quantitative evidence of enhanced interpretability, grounding …

Error-Driven Density Control For Compact Gaussian Splatting Under Sparse Supervision, Abdelrhman Elrawy

Theses and Dissertations (Comprehensive)

This thesis studies efficiency and stability challenges in Gaussian-splatting-based reconstruction under sparse supervision. In few-shot novel view synthesis, standard 3D Gaussian Splatting (3DGS) can overfit the limited training views and grow an unnecessarily large number of primitives due to limitations in its Adaptive Density Control (ADC) mechanism. This thesis introduces an error-driven reformulation of ADC that triggers densification using opacity gradients as a lightweight proxy for rendering error, and shows that such aggressive densification must be paired with delayed and conservative pruning to prevent destructive create--destroy cycles. When combined with depth-based geometric regularization, the resulting framework produces substantially more compact …

Differential Impact Of Admission Type And Clinical Complexity On Diabetes Hospitalization Costs Among African American And Hispanic Patients In Southeastern Virginia, Ismail El Moudden, Asra Amidi, Reem Sharaf-Alddin, Michael C. Bittner, Qi Zhang

Department of Obstetrics & Gynecology Faculty Publications

Background

Diabetes mellitus (DM) imposes substantial healthcare costs with documented disparities among African Americans and Hispanic patients. To inform care delivery and resource allocation, this study identified hospitalization cost predictors among African American and Hispanic patients with diabetes in Southeastern Virginia.

Methods

We analyzed 6,011 hospital discharges from the Virginia Health Information database (2016-2020) for adults aged 18-85 with diabetes. Discharges were classified by Medicare Severity Diagnosis-Related Groups: DM with complications/comorbidities (DCC, n = 3,328), DM with major complications/comorbidities (DMCC, n = 1,518), and DM without major complications/comorbidities (DWO, n = 1,165). Because cost distributions were right-skewed (skewness 3.5-8.24), we …

Limits Of Machine Learning For Predicting Stillbirth At Term: A Population-Based Cohort Of 302,192 Pregnancies In Northern Italy, M. Benuzzi, T. Kawakita, F. Monari, B. Melis, G. Guariglia, A. La Marca, I. Neri, G. Saade, F. Facchinetti, I. El Moudden

Department of Obstetrics & Gynecology Faculty Publications

Background

Objective

Study Design

Modeling Healthcare Data With Logistic Quantile And Uniform-Based Mixture Polynomial Distributions, Mohan D. Pant, Aditya Chakraborty, Jovanna A. Tracz

Epidemiology, Biostatistics, & Environmental Health Faculty Publications

Continuous healthcare data often deviate from normality, which can substantially increase the risk of making invalid inferences, given that many inferential statistical procedures rely on normality assumption. To obviate this issue, we propose a new family of non-normal distributions based on a linear combination of the quantile functions of standard logistic and uniform (0, 1) distributions. This new family of non-normal distributions is studied within three different methods: L-moments, conventional moments, and percentiles. Its performance is compared among the three methods in the context of parameter estimation and data modeling. The results of Monte Carlo simulation and bootstrapping techniques indicate …

Understanding Cancer Information Burden And Quality Concerns: The Roles Of Medical Statistics Literacy, Online Health Information Confidence, And Social Media, Aditya Chakraborty, Sikta Das, Amy Myat Phyo

Epidemiology, Biostatistics, & Environmental Health Faculty Publications

Understanding factors shaping perceptions of cancer-related information is important for public health communication. This study examined associations of three information-related predictors, understanding medical statistics, confidence in online health information, and difficulty interpreting social media information, with two outcomes: cancer information was too hard to understand (CTU) and concern about cancer information quality (CCQ). Data came from the 2022 Health Information National Trends Survey. The cross-sectional study included 1972 participants. Survey-weighted univariable and multivariable logistic regression models assessed associations between the three predictors and CTU and CCQ, adjusting for sociodemographic characteristics, health insurance, and cancer history. LASSO logistic regression, random forest, …

Technological Interventions For Reducing Climate Change Impacts On Health: An Overview Of Future Possibilities, Sujatha Alla, Vijay Kumar Chattu, Bawa Singh

Engineering Management & Systems Engineering Faculty Publications

Globally, over five million deaths annually are attributed to extreme weather events exacerbated by climate change, such as heatwaves, hurricanes, wildfires, droughts, and floods, which create health crises and economic losses. The 2015 Paris Declaration emphasized climate technology mechanisms, including research and development, to enhance resilience and reduce greenhouse gas emissions. Biotechnological advancements have mitigated about 20% of economic losses since the 1960s, highlighting technology’s potential to address climate-induced health risks.

Validity Assessment Of Resting Heart Rate Variability From The Garmin Health Snapshot, Kayla M. Porter, Andrew Flatt

Honors College Theses

Purpose: To assess the agreement between the Garmin Forerunner 265 (a commercially available sports watch) and a single-channel electrocardiographic (ECG) chest strap for determining resting heart rate variability (HRV). Secondary aims were to assess the impact of skin tone and body position on measurement accuracy.

Methods: Young adults (n = 30, 57% women) aged 18–39 years without known cardiovascular conditions and without tattooing or scarring on the dorsal left wrist were recruited. HRV was recorded simultaneously using ECG and the Forerunner 265’s optical sensor during Garmin’s 2-minute “Health Snapshot.” Measurements were obtained in three standardized positions: supine, seated, and standing. …

A Novel Mathematical Model Of Hiv Transmission Incorporating The Effects Of Treatment And Pre-Exposure Prophylaxis: Sensitivity Analysis And Numerical Simulations, Erick Manuel Delgado Moya

Mathematical Modelling and Numerical Simulation with Applications

Human immunodeficiency virus (HIV) continues to be a public health problem in many countries of the world, and Pre-exposure prophylaxis (PrEP) is a preventive method for HIV, which has shown great efficacy and is in use worldwide. This work presents a new mathematical model for HIV transmission incorporating PrEP use and evaluates the impact of PrEP along with its increasing use in a population. The construction of the model takes into account three forms of diagnosis: diagnosis of individuals in risky sexual contact, diagnosis after risky contact (diagnosis in the undiagnosed infected compartment), and diagnosis associated with attempting to enter …

Analysis And Numerical Investigation Of A Breast Cancer Treatment Model Incorporating Ketogenic Diet And Immune Boosters Via Optimal Control Theory, Kunnisai Muniroh, Ummu Habibah, Wuryansari Muharini Kusumawinahyu, Nur’Izzati Hamdan

Mathematical Modelling and Numerical Simulation with Applications

This paper develops a mathematical model to investigate breast cancer dynamics by incorporating tumor–immune interactions, ketogenic diet effects, and medical treatment. The model is formulated as a system of nonlinear ordinary differential equations and analyzed within an optimal control framework. Time-dependent control variables are introduced to represent treatment strategies aimed at minimizing tumor progression while reducing therapeutic costs. The model’s well-posedness is established through positivity and boundedness analysis. The necessary conditions for optimality are derived using Pontryagin’s Minimum Principle, resulting in a coupled system of state and adjoint equations. Numerical solutions are obtained using the fourth-order Runge–Kutta method combined with …

Mathematical Analysis And Numerical Simulation Of A Fractional-Order Sir-Si Model For Malaria Transmission Dynamics, Gassan A. M. O. Farah, Abdulaziz Y. A. Mukhtar, Kailash C. Patidar

Mathematical Modelling and Numerical Simulation with Applications

This study investigates the complex transmission dynamics of malaria, a critical global health challenge, with a focus on the African continent. We introduce a novel approach that employs Fractional Differential Equations (FDEs) to advance the understanding of malaria spread and control. Specifically, we develop a new SIR-SI model using the Caputo fractional operator, which captures the memory effects and time-delay characteristics inherent in real-world epidemiological systems. A detailed analysis of the model's solvability and uniqueness is conducted using fixed-point theory. To obtain an analytical solution, the system is solved via the Laplace transform method, with solutions expressed in closed form …