Digital Commons Network^™

Advancing Multi-Modal Learning, Swetha Sirnam

Graduate Studies Theses and Dissertations 2026

Multimodal learning aims to develop systems that understand and reason across multiple sources of information such as images, videos, audio, and text. These capabilities are central to applications including visual search, digital assistants, robotics, and visual question answering. Despite recent progress in multimodal large language models, current approaches struggle to capture the structured complexity of real-world visual data. Fine-grained visual details are often lost during multimodal alignment, models frequently fail to capture nuanced human intent, and existing methods remain limited in understanding long-form videos with complex temporal dynamics. Furthermore, many evaluation benchmarks rely heavily on explicit visual cues, which can …

Multimodal Representation Learning For Face Understanding: From Caption Supervision To Foundation Model Adaptation, Md Mahedi Hasan

Graduate Theses, Dissertations, and Problem Reports (ETD)

The rapid advancement of intelligent surveillance systems and the increasing demand for reliable biometric identification in border security, public safety, and digital forensics require robust face understanding under unconstrained conditions, including low resolution, pose variation, and occlusion. While Vision Transformer (ViT)-based foundation models have greatly improved visual representation learning, their patch-based tokenization and lack of spatial inductive bias limit their ability to capture fine-grained details in low-resolution inputs. This dissertation investigates multimodal representation learning for face understanding through natural language supervision, large-scale face-caption pre-training, and parameter-efficient foundation model adaptation. It hypothesizes that textual supervision provides complementary semantic cues that improve …

Use Of Generative Ai In Scholarly Research: Challenges And Opportunities, Marco Marabelli, Robert M. Davison, Giovanni Gatti, Ankita Srivastava, Monideepa Tarafdar

Computer Information Systems Faculty Publications

The focus of this panel report is to present evidence of a panel discussion that took place in October 2025 on the use of generative artificial intelligence (GAI) in academic research and publishing. The panel included publishers (Elsevier, Emerald, Gailot Press, MIT Press, Nature/Springer, Wiley) and editors in chief (representing the Information Systems Journal, Journal of the AIS, Journal of Data Science and ICES Journal of Marine Science). Key outcomes of the panel are the following: 1) publishers and editors have some disagreements regarding GAI use; 2) besides disagreements, all recognize that GAI is here to stay, therefore outstanding issues …

Large Language Model-Assisted Research Question Development In Public Health: A Case Study In The Special Supplemental Nutrition Program For Women, Infants, And Children (Wic), Qi Zhang, Bidusha Neupane, Priyanka Patel, Futun N. Alkhalifah, Yi He, Leslie Hodges

Health Behavior, Policy & Management Faculty Publications

Objective:

To assess the feasibility of using large language models (LLMs) to develop research questions about changes to the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) food packages.

Design:

We conducted a controlled experiment using ChatGPT-4 and its plugin, MixerBox Scholarly, to generate research questions based on a section of the USDA summary of the final public comments on the WIC revision. Five questions weekly for three weeks were generated using LLMs under two conditions: fed with or without relevant literature. The experiment generated 90 questions, which were evaluated using the FINER criteria (Feasibility, Innovation, Novelty, Ethics, …

Fleet: A Federated Learning Emulation And Evaluation Testbed For Holistic Research, Osama Abu Hamdan, Hao Che, Engin Arslan, Md Arifuzzaman

Computer Science Faculty Research & Creative Works

Federated Learning (FL) presents a robust paradigm for privacy-preserving, decentralized machine learning. However, a significant gap persists between the theoretical design of FL algorithms and their practical performance, largely because existing evaluation tools often fail to model realistic operational conditions. Many testbeds oversimplify the critical dynamics among algorithmic efficiency, client-level heterogeneity, and continuously evolving network infrastructure. To address this challenge, we introduce the Federated Learning Emulation and Evaluation Testbed (FLEET). This comprehensive platform provides a scalable and configurable environment by integrating a versatile, framework-agnostic learning component with a high-fidelity network emulator. FLEET supports diverse machine learning frameworks, customizable real-world network …

Explaining The Unseen: Multimodal Vision-Language Reasoning For Situational Awareness In Underground Mining Disasters, Mizanur Rahman Jewel, Mohamed Elmahallawy, Sanjay Kumar Madria, Samuel Frimpong

Computer Science Faculty Research & Creative Works

Underground mining disasters produce pervasive darkness, dust, and collapses that obscure vision and make situational awareness difficult for humans and conventional systems. To address this, we propose MDSE, Multimodal Disaster Situation Explainer, a novel vision-language framework that automatically generates detailed textual explanations of post-disaster underground scenes. MDSE has three-fold innovations: (i) Context-Aware Cross-Attention for robust alignment of visual and textual features even under severe degradation; (ii) Segmentation-aware dual pathway visual encoding that fuses global and region-specific embeddings; and (iii) Resource-Efficient Transformer-Based Language Model for expressive caption generation with minimal compute cost. To support this task, we present the Underground Mine …

Using Inferences From Natural Language Narratives To Improve Video Question Answering, Hoang Bao Truong

Graduate Studies Theses and Dissertations 2026

The purpose of this thesis is to examine how structured narrative representation and reasoning strategies jointly affect the performance of Video Question Answering (VideoQA). Although recent advances in large multimodal and language models have improved video description and retrieval, causal and counterfactual reasoning remain challenging. This thesis studies an inference-only pipeline in which videos are converted into natural-language narratives, and those narratives may optionally be transformed into a structured NarrativeML format before multiple-choice questions are answered without task-specific fine-tuning. NarrativeML is a machine-readable narrative representation that explicitly encodes entities, events, and their causal and temporal relations. Three representational contexts are …

Thinkmatter: Panoramic-Aware Instructional Semantics For Monocular Vision-And-Language Navigation, Guangzhao Dai, Shuo Wang, Hao Zhao, Bin Zhu, Qianru Sun, Xiangbo Shu

Research Collection School Of Computing and Information Systems

Vision-and-Language Navigation in continuous environments (VLN-CE) requires an embodied robot to navigate the target destination following the natural language instruction. Most existing methods use panoramic RGB-D cameras for 360° observation of environments. However, these methods struggle in real-world applications because of the higher cost of panoramic RGB-D cameras. This paper studies a low-cost and practical VLN-CE setting, e.g., using monocular cameras of limited field of view, which means “Look Less” for visual observations and environment semantics. In this paper, we propose a ThinkMatter framework for monocular VLN-CE, where we motivate monocular robots to “Think More” by 1) generating novel views …

Purified Zero-Shot Sketch-Based Image Retrieval, Yang Zhou, Jingru Yang, Jin Wang, Kaixiang Huang, Guodong Lu, Shengfeng He

Research Collection School Of Computing and Information Systems

Sketches, as a new solution in multimedia systems that can replace natural language, are characterized by sparse visual cues such as simple strokes that differ significantly from natural images containing complex elements such as background, foreground, and texture. This misalignment poses substantial challenges for zero-shot sketch-based image retrieval (ZS-SBIR). Prior approaches match sketches to full images and tend to overlook redundant elements in natural images, leading to model distraction and semantic ambiguity. To address this issue, we introduce a distraction-agnostic framework, purified cross-domain matching (PuXIM), which operates on a straightforward principle: masking and matching. We devise a visual-cross-linguistic (VxL) sampler …

Patterns Of Llm Weaponization: A Comparative Analysis Of Exploitation Incidents Across Commercial Ai Systems, George Antoniou

Faculty and Staff Publications & Presentations

This comparative study examines patterns of Large Language Model (LLM) weaponization through systematic analysis of four major exploitation incidents spanning from 2023-2025. While existing research focuses on isolated incidents or theoretical vulnerabilities, this study provides one of the first comprehensive comparative frameworks analyzing exploitation patterns across state-sponsored cyber-espionage (Anthropic Claude incident), academic security research (GPT- 4 autonomous privilege escalation), social engineering platforms (SpearBot phishing framework), and underground criminal commoditization (WormGPT/FraudGPT ecosystem). Through comparative analysis across eight dimensions: Adversary sophistication, target selection, exploitation techniques, autonomy levels, detection evasion, attribution challenges, defensive gaps, and capability democratization, this research identifies critical cross-case patterns …

Supply Chain Network Based On Blockchain And Intelligent Agent, Hiba Hamdi Hassan, Rana Fareed Ghani

Journal of Soft Computing and Computer Applications

In agricultural supply chains, the complexity and indeterminacy pose serious challenges to traceability, reliability and confidence today. This challenge is especially acute in the olive oil industry where adulteration, wrong labeling, and uneven chemical quality threaten the actual well-being of producers and consumers. The project aims to design a blockchain-based hybrid architecture with intelligent agents (FNNs) to enhance transparency, reliability and responsiveness in the olive oil supply chain. The Blockchain component enables a completely open, tamper-proof ledger to be built in a very decentralized way and preserved as an archive of every account of its transactions. The intelligent agents contribute …

Intelligent Extensible Markup Language Encryption Using Type-2 Fuzzy Logic, Faiez Musa Lahmood Alrufaye, Seham Ahmed Hashem

Journal of Soft Computing and Computer Applications

Financial and commercial institutions increasingly rely on Extensible Markup Language (XML) files as a standard means of exchanging data. However, this extensive use has created serious security challenges due to the fact that these files contain sensitive information such as bank card numbers and expiration dates. Relying on traditional full file encryption methods achieves a high degree of security, but it causes problems related to the large file sizes that consume memory and the long encryption and decryption times, which reduces the efficiency of systems when dealing with a large number of daily transactions. Methods based on Type-1 Fuzzy Logic …

Hate Speech Detection Using Optimized Feature Representation Via Spiral-Grey Wolf Optimizer-Based Machine Learning Approaches, Noor S. Farhan, Matheel E. Abdulmunim, Hasanen S. Abdullah

Journal of Soft Computing and Computer Applications

Hate speech detection is crucial as social media diversifies. This research present a lightweight, scalable system using traditional machine learning methods along with a new approach called Spiral-Grey Wolf Optimizer (S-GWO).

S-GWO effectively selects key features that consider both meaning and content from the Term Frequency Inverse Document Frequency (TF-IDF) space, leading to high-quality representation without excessive computing power.

The propoused system was tested on Arabic and another English datasets using six machine learning methods: SVM, RF, LR, KNN, NB, and SGD. It achieved 92% accuracy and F1 score on the Arabic dataset, while reaching 100% accuracy on the English …

A Method Of Heuristic Human-Llm Collaborative Source Search, Yi Chen, Sihang Qiu, Zhengqiu Zhu, Yatai Ji, Yong Zhao, Rusheng Ju

Journal of System Simulation

Abstract: Traditional source search algorithms are prone to local optimization, and source search methods combining crowdsourcing and human-AI collaboration suffer from low cost-efficiency due to human intervention. In this study, we proposed a lightweight human-AI collaboration framework that utilized multi-modal large language models (MLLMs) to achieve visual-language conversion, combined chain-of-thought (CoT) reasoning to optimize decision-making, and constructed a heuristic strategy that incorporated probability distribution filtering and a balance between exploitation and exploration. The effectiveness of the framework was verified by experiments. The human-AI alignment heuristic strategy with large language model adaptation design provides a new idea to reduce manual …

Advancements In Perfect Matchings Within Neutrosophic Fuzzy Graphs: Theory And Applications, Muhammad Saeed, Fatima Razaq

Neutrosophic Systems with Applications

Graph theory has been widely used in exemplifying relational structure, and a greater generalization to fuzzy and neutrosophic space enables the exemplification of uncertainty, indeterminacy, and inconsistency in complex systems. Some of these extensions include the neutrosophic fuzzy graphs that provide a more detailed description of the loose relations between the vertices and the edges. However, unlike in classical graph theory, where the concepts of matching and perfect matching are well developed, very little has been studied on how the two concepts can be extended to neutrosophic fuzzy graphs. To seal this gap, the current paper develops and defines the …

Integrating Mcdm Techniques For Optimized Task Offloading In Multi-Uav- Enaled Mobile Edge Computing System, Amira Salam

Neutrosophic Systems with Applications

Carefully choosing a task offloading strategy is crucial for optimizing task scheduling and offloading strategies in a multi-UAV system. Regarding cost, responsiveness, scalability, and data security, each strategy has pros and cons. As a result, selecting the best offloading technique is essential to multi-UAV mobile edge computing task scheduling optimization. A methodical and well-informed decision-making process accomplishes this.

The current study introduces a new hybrid methodology for the multi-criteria decision-making (MCDM) model, known as IVNs-DEMATEL-ANP-VIKOR. The purpose of this model is to find and choose an efficient task offloading strategy. With this method, we use IVNs-DEMATEL to show how different …

Tests Without Borders: A Global Approach To Measuring Visualization Literacy, Olivia A. Guess

McKelvey School of Engineering Graduate Student Theses & Dissertations

Visualization literacy assessments shape how we understand people's ability to interpret data, yet most existing instruments embed Western datasets and assumptions that limit their relevance for global audiences. This thesis argues that because data is personal, assessments must also be culturally grounded. We introduce a unified framework for adapting the Mini-VLAT into 22 regionally responsive short-form assessments, each retaining the structure of the original test while incorporating datasets and scenarios tailored to specific regions around the world. To demonstrate how such adaptations can be customized and validated, we present a detailed case study of a Ghana-adapted Mini-VLAT, developed in collaboration …

Ankimedbench: Evaluating Hierarchical Medical Knowledge In Language Model Embeddings, Neel Patel

UNLV Theses, Dissertations, Professional Papers, and Capstones

Despite achieving over 90% accuracy on medical benchmarks, recent studies show physicians cannot effectively leverage language models to improve clinical reasoning. Current benchmarks test isolated factual recall, but clinical practice requires hierarchical navigation through diagnostic categories—starting broad and narrowing systematically from chest pain to cardiovascular pathology to myocardial infarction to specific STEMI types. Existing evaluations cannot measure whether models preserve this taxonomic structure essential for clinical reasoning.

We introduce AnkiMedBench, built from 16,512 medical flashcards used by students preparing for licensing exams. Cards are organized across six hierarchy levels spanning 16 broad medical specialties to 672 specific diseases and conditions. …

Achieving More Accurate And Interpretable Fraud Detection With Double Machine Learning, Jeremy Andrew Berry

Theses and Dissertations

Fraud detection remains a critical challenge across industries such as insurance, healthcare, finance, and government. Global losses from fraud and financial crime are estimated in the trillions annually, including billions in healthcare and insurance fraud alone. While effective for prediction, traditional machine learning methods often lack causal interpretability and struggle to adapt to evolving fraud tactics. This dissertation investigates the application of Double Machine Learning (DML), an emerging causal inference technique, to enhance both the accuracy and interpretability of fraud analytics. The research compares DML against established causal inference approaches, leveraging a meta-learning framework to evaluate model performance on accuracy, …

Cross-Domain Transfer Learning Of Tabular Data With Heterogeneous Feature Spaces, Kazi Fuad Bin Akhter

Tennessee State University Alumni Theses and Dissertations

Transfer learning is a cornerstone of artificial intelligence (AI), enabling the learning of new domains with limited labeled data by reusing knowledge from large-scale foundation models. Transfer learning has achieved remarkable success with vision and language models due to the homogeneity in image and text data. In contrast, a heterogeneous feature space, structured in rows and columns as tabular data, remains underexplored in the transfer-learning literature. The inductive bias of tabular data with heterogeneous feature spaces from disparate application domains is very different from image and text data, which complicates transfer learning. Several recent studies have attempted within-domain and limited …

Prism (Proxy Recognition And Inclusion Scoring Method), Destiny Raburnel, Crystal Tubbs, Md Abdullah Al Hafiz Khan

Symposium of Student Scholars

AI-driven automated hiring tools are reshaping how companies find talent, but they often reproduce the hidden biases embedded in their training data. Our project, PRISM (Proxy Recognition and Inclusion Scoring Method), investigates how subtle demographic signals, specifically first names associated with gender and race, influence AI resume screening even when candidates have identical qualifications. We built a controlled dataset of resumes that are identical in every way except for the applicant's first name, with each resume using a racially neutral surname to isolate how first names alone affect scoring. We tested these resumes against job postings in technology, healthcare, and …

A Proposed Study Of Tone Indicators In Sentimental Analysis And Emotion Detection, Andrea Llanas

Infinite Loop

Sentimental analysis and emotion detection have been an ever-growing field in academic literature in recent years [1,2,3]. There are many methods and techniques to distinguish positive and negative tokens as well as classification of emotions respectively. However, the use of tone indicators has been relatively underexplored within the field.

Tone indicators are a relatively recent trend in social media. Users denote a positive or negative connotation as well as an emotion in a sentence at the moment of conception with syntax such as “/s,” “/pos,” and “/neg.” These annotations often are context-free, or do not depend on previously declared information, …

Understanding Medical Information And Emotional Support Needs In Mental Health Questions With Large Language Models, Chen Liu, William Yu Chung Wang, Gohar Khan

All Works

Purpose – This study seeks to bridge the gap between users’ multidimensional needs and the single-task capabilities of existing Mental Health Question Answering (MHQA) systems by tackling the underexplored challenge of jointly understanding medical informational needs and emotional support needs within complex consumer mental health inquiries. Design/methodology/approach – Grounded in Rhetorical Structure Theory (RST), the proposed Multi-Needs and Context Recognition (MNCR) framework decomposes mental health question understanding task into four interrelated subtasks: Medical Needs Recognition (MNR), Medical Needs-related Context Extraction (MNCE), Emotional Needs Recognition (ENR) and Emotional Needs-related Context Extraction (ENCE). A new benchmark dataset, MHQ-MedEmo, was constructed through multi-layered …

Bridging Machine Learning And Islamic Scholarship: A Study In Hadith Translation And Similarity Analysis, Asiyah R. Speight

Student Scholar Symposium Abstracts and Posters

Translation of Islamic religious texts poses unique challenges requiring both linguistic and theological expertise. This study explores the application of neural machine translation (NMT) models to Arabic-English hadith translation while analyzing semantic similarity patterns across different human translations. Using the complete Sahih Bukhari corpus (7,550 hadiths) as the primary dataset, we adopt a dual approach combining transfer learning and comprehensive neural network analysis to demonstrate the critical impact of corpus size on model performance.

First, we fine-tune a pre-trained MarianMT Arabic-English translation model on the full Sahih Bukhari corpus, comparing models trained on 40 hadiths versus 7,550 hadiths. Performance is …

Large Language Models As Machines Of Beauty: Cognitive Averaging, Latent Space Geometry, And The Entropic Foundations Of Aesthetic Preference, Daniel Plate, James Hutson

Faculty Scholarship

This study advances the position that large language models (LLMs) and human perceptual systems are governed by a shared computational drive toward prototypicality, entropy reduction, and aesthetic coherence. Drawing on developmental evidence that infants exhibit early preferences for facial symmetry and averageness, the analysis situates aesthetic preference within broader research on processing fluency and predictive coding, emphasizing that biological perception rewards stimuli that reduce uncertainty and support efficient information compression. This foundation is used to examine how LLMs, through cross-entropy optimization, perplexity minimization, and latent space clustering, converge on high-density representational regions that operate as statistical prototypes of linguistic and …

Proverag: Provenance-Driven Vulnerability Analysis With Automated Retrieval-Augmented Llms, Reza Fayyazi, Stella Hoyos Trueba, Michael Zuzak, Jay Yang

Institute for Informatics and Applied Technology Scholarship

In cybersecurity, security analysts constantly face the challenge of mitigating newly discovered vulnerabilities in real-time, with over 300,000 vulnerabilities identified since 1999. The sheer volume of known vulnerabilities complicates the detection of patterns for unknown threats. While LLMs can assist, they often hallucinate and lack alignment with recent threats. Over 40,000 vulnerabilities have been identified in 2024 alone, which are introduced after most popular LLMs’ (e.g., GPT-5) training data cutoff. This raises a major challenge of leveraging LLMs in cybersecurity, where accuracy and up-to-date information are paramount. Therefore, we aim to improve the adaptation of LLMs in vulnerability analysis by …

Digital Communications Between Firms And Investors: Impact Of Explanatory Responses On Investor Engagement In Online Financial Q&A, Runyu Wang, Zili Zhang, Keng Siau, Ziqiong Zhang

Research Collection School Of Computing and Information Systems

The emerging trend of digital communications between firms and investors through online question-and-answer (Q&A) platforms is recognized as a vital strategy for managing investor relations, contributing to enhanced market efficiency and information transparency through increased information exchange. Potential investors can seek responses from firm managers to address their information needs, thereby mitigating market uncertainties. To provide foundational insights, we conduct a survey of investors to assess their awareness, usage, and perceptions of firm-investor Q&A platforms. In the subsequent empirical study, we specifically focus on the substance of managers’ responses, which are primarily aimed at clarifying firm events or information. In …

Emotion Analysis And Neural Language Models For Classification, Andrew Mackey

Graduate Theses and Dissertations

Emotion analysis is a branch of artificial intelligence and natural language processing focused on recognizing emotions hidden throughout various forms of digital data, including text, images, and multi-modal representations. In this dissertation, we present four published and planned works that investigate different methodologies for natural language analysis tasks using deep learning techniques. The first published work we present considers the task of identifying fake news using various text and emotion representations. We demonstrate that emotion representations combined with word embedding techniques can improve the accuracy of fake news classification. Our second published work further investigates the fake news classification task …

Llm-Assisted Cwe Identification, Severity Assessment, And Vulnerability Description Generation, Mohammad Jalili Torkamani

School of Computing: Dissertations, Theses, and Student Research

Identifying the underlying weakness types and assessing their severity using CWE and CVSS standards are critical steps in software vulnerability management. While timely assessment of vulnerabilities mitigates the impact of severe security incidents, automating joint CWE identification and severity assessment remains challenging due to the heterogeneity of vulnerabilities across different code granularities and programming languages. In addition, generating vulnerability descriptions is often time-consuming, as it requires extensive manual review, validation, and writing by security experts.

In this thesis, we leverage the capabilities of Large Language Models (LLMs) to automate the identification of CWE identifiers and the assessment of their severity …

Enhancing Ad/Adrd Management Through Ihelpcare: A Compliant And Culturally Sensitive Ai-Driven Digital Healthcare Platform, Trisha Bhowmick

Master's Theses

The digital healthcare field is expanding fast, and now it requires platforms that use advanced technology and maintain robust data security and compliance practices. In the present paper, we present the main structure, key methods, and compliance strategies of the digital healthcare system iHelpCare, which, while fully meeting the HIPAA/GDPR requirements, provides health services more accessible, efficient, and inclusive. The proposed platform is powered by AI for personalized care solutions, with the main emphasis on preventive health management and providing tools for people with disabilities.

iHelpCare achieves real-time patient monitoring while securing medical data management and easy communication between patients, …