Digital Commons Network^™

Development And Evaluation Of Machine Learning Models For Early Pediatric Sepsis Prediction, Ancita M. Andrade

Browse all Theses and Dissertations

Sepsis is a leading cause of pediatric mortality, claiming more lives in the United States annually than all childhood cancers combined. Early identification in Emergency Departments (EDs) remains challenging, as the current Phoenix criteria establishes an updated international consensus definition for sepsis, however is not designed for use as a screening tool. This study aimed to develop predictive models identifying pediatric patients at risk of sepsis within 24 hours of admission. Multiple tree-based and deep learning models were trained utilizing clinical and laboratory data from the initial four hours of presentation. Both the LightGBM and LSTM architectures demonstrated superior performance, …

Wearable Sensor Data Analysis For Machine Learning-Based Detection Of Posture And Autonomic Responses, Chaitanya Vardhini Anumula

Browse all Theses and Dissertations

This study investigates how Iyengar yoga postures influence autonomic nervous system (ANS) activity by analyzing multimodal physiological signals collected via wearable sensors. The physiological mechanisms underlying Iyengar yoga’s therapeutic effects remain under-explored at the granular, pose-level. Using data collected from 16 participants, this research evaluates whether machine learning models can distinguish between baseline, parasympathetic-dominant, and sympathetic-dominant states based on wrist-worn sensor data. The goals were to explore whether subtle postural variations elicit measurable autonomic responses and to identify which sensor features most effectively capture these changes. Participants performed a sequence of yoga poses while wearing synchronized sensors measuring electrodermal activity …

Impact Of Graph Structures For Rag Outcomes In Llms, Chris Davis Jaldi

Browse all Theses and Dissertations

Explainability, interpretability and adaptability (EIA) remain three central motivations for next-generation Artificial Intelligence (AI), especially as Large Language Models (LLMs) continue to engage with ever-increasing knowledge bodies. As the landscape pushes toward controllable agentic Retrieval-Augmented Generation (RAG) systems where AI agents engage in iterative, guided reasoning, a critical question arises as to the extent to which the knowledge design itself shapes these models' reasoning behavior. This work conducts a systematic evaluation of how different conceptualizations and representation of the identical knowledge affect an LLM's path-based reasoning capabilities. Through the introduction of controlled variations along graph structural complexity, linguistic and semantic …

Re-Parameterizing Adversarial Reprogramming In Low-Dimensional Subspace For Efficient Software Vulnerability Detection, Hootan Alavizadeh

Browse all Theses and Dissertations

Software vulnerabilities are a major cause of security breaches, making effective detection critical. Traditional learning-based methods require large datasets and significant computational resources, which are often impractical due to high annotation costs and data scarcity. To address this, we propose an innovative system, RearVul, which Re-parameterizes adversarial reprogramming in a low-dimensional subspace for software vulnerability detection. Unlike conventional approaches, RearVul repurposes a pre-trained classification model using adversarial reprogramming, enabling detection with minimal modifications. It learns a universal perturbation applied to program representations, preserving the original model’s feature extraction capabilities while adapting it to a new domain. Furthermore, we introduce a …

Subjective Readiness Forecasting Using Supervised Machine Learning And Wearable Device Data, Nathaniel Michael Weiland

Browse all Theses and Dissertations

Recent advances in wearable technology allow continuous monitoring of physiological and behavioral data, opening new opportunities for real-time assessments of readiness and well-being. However, creating predictive models that generalize across diverse users remains challenging, especially in high-stakes settings like the military, where preventable injuries, illnesses, and stress-related performance declines are frequent. This research assesses the feasibility of using supervised machine learning models trained on wearable device data to predict subjective readiness indicators—recovery, stress, injury, and illness. Data from over 10,000 users in the OHWS (Optimizing the Human Weapons System) program combined daily check ins with physiological metrics from Garmin, Polar, …

Dataset Generation For Routing Policy Study In Ad Hoc Wireless Networks, Vishnu Vishnu Priya

Browse all Theses and Dissertations

Ad Hoc wireless networks, with their decentralized architecture and dynamic topology, present challenges in reliable and energy-efficient routing. While machine learning (ML) and reinforcement learning (RL) offer promising solutions, progress is limited by the lack of realistic, high-fidelity datasets. This research introduces a simulation-based framework for generating four diverse datasets representing combinations of node mobility (mobile vs. static) and spatial distribution (random vs. clustered). Each dataset captures critical metrics such as Signal-to-Interference-plus-Noise Ratio (SINR), bottleneck rate, and power consumption across multi-hop paths. A lookahead-based greedy routing algorithm with scenario-aware power control is implemented to emulate practical behavior. Supervised ML models, …

Isometric Centroid Encoder (Ice) And Synthetic Data Generation Approaches For Biological Datasets, Prathyusha Kanakamalla

Browse all Theses and Dissertations

This thesis addressed two main challenges in biological data analysis: structure-preserving dimensionality reduction and synthetic data generation for small sample datasets. I proposed the Isometric Centroid Encoder (ICE), a supervised dimensionality reduction method that preserves pairwise distances between class centroids during dimension reduction. Unlike existing methods like Centroid Encoder and Super Encoder, ICE explicitly maintains geometric relationships between biological classes, achieving nearly perfect structure preservation at C dimensions (where C equals the number of classes) with strong performance even in 2D and 3D spaces. Additionally, I compared three generative models (VAE, LSH-GAN, and scDiffusion) for synthetic data generation on small …

Ultrasonic Sensor-Based Sound Synthesis Using Raspberry Pi Pico W, Niraj Jaishwal

Williams Honors College, Honors Research Projects

At the intersection of Human Computer Interaction and digital art, this project transforms simple motion into musical expression. It explores an interactive real-time sound synthesis system using ultrasonic sensors to generate continuous audio. The objective is to design a system that maps physical distances into musical parameters such as pitch and amplitude, which will create a responsive audio environment. Two ultrasonic sensors are used in combination with the Raspberry Pi Pico W microcontroller running CircuitPython and Adafruit Audio Hat for real-time sound output. One sensor controls the pitch of the generated tone, while the other controls volume. This enables expressive …

Personalized Physics Learning Through Ai: Insights From Problem Generation, Chatbot Dialogues, And Intelligent Tutoring Systems, Atharva Dange

Physics Dissertations - Archive

Artificial intelligence (AI) is poised to transform science education, yet questions remain on how best to integrate these technologies into teaching and learning. This dissertation investigates the use of AI-driven tools in university physics courses through three complementary studies. In the first study, a generative language model (ChatGPT) was used to create novel physics homework problems aligned with course objectives. Analysis showed that, after expert vetting, AI-generated questions can foster higher-order problem-solving and reduce student reliance on solution memorization, though careful instructor oversight is required to ensure accuracy. The second study embedded an AI chatbot as a learning aid in …

Llm-Driven Fmea For Safe Human-Robot Collaboration In Disassembly, Morteza Jalali Alenjareghi, Samira Keivanpour, Yuvin Adnarain Chinniah, Sabrina Jocelyn

Articles dans des actes de congrès

Disassembly operations often present unstructured and unpredictable scenarios, such as handling hazardous materials, addressing ergonomic strain, and managing dynamic robot interactions that pose safety risks. To tackle these challenges, we propose an innovative use of large language models (LLMs) to enhance failure mode and effect analysis (FMEA) in the context of human-robot collaboration (HRC) for disassembly tasks. We developed an LLM system leveraging retrieval-augmented generation (RAG) for real-time risk analysis and recommendation generation. RAG retrieves domain-specific information from the FMEA knowledge database, enabling accurate risk analysis, contextual understanding, and relevant recommendations based on user input and operational data. Evaluation of …

Leveraging Counterfactuals For Enhanced Natural Language Inference: A Multitask Knowledge Distillation Approach, Brian J. Davis

Browse all Theses and Dissertations

Natural-language inference (NLI) asks whether a hypothesis is entailed by, contradicts, or is neutral with respect to a premise. Modern transformers reach high raw accuracy on benchmarks such as SNLI, MNLI, and ANLI, yet they often rely on brittle lexical shortcuts and provide little insight into their decision process. This thesis shows that counterfactual-augmented knowledge distillation can simultaneously boost robustness and supply faithful, token-level explanations—without scaling model size. Four T5-v1_1 students (60M, 220M, 770M, 3B parameters) are trained under four curricula: (1) standard fine-tuning, (2) fine-tuning with free-text rationales, (3) multi-task distillation with naive counterfactuals, and (4) multi-task distillation with …

Optimizing Cloud Computing Resources For Operational Cost And Application Performance Using Machine Learning, Isaac K. Matthew

Browse all Theses and Dissertations

As AI-driven workloads accelerate the growth of cloud initiatives and spending, resource waste also increases due to persistent inefficiencies in cloud compute and infrastructure management. Overprovisioned resources and suboptimal configurations often lead to operational inefficiencies and unnecessary financial overhead. These challenges arise from the difficulty of anticipating resource demands in dynamic workloads and selecting suitable virtual machines to ensure optimal performance. Our research proposes a holistic, data-driven framework for managing cloud compute resources that reduces costs without compromising application performance. We integrate a predictive, model-driven, threshold-based autoscaling solution for cloud-native applications with an optimized instance right-sizing approach to select cost-effective …

Reducing Operator Training Time Through Virtual Reality: A Case Study On The Lpkf Protomat E44 Machine, Joshua C. Patel

Browse all Theses and Dissertations

This thesis presents the development of an immersive virtual reality (VR) simulation that replicates the operation of the LPKF ProtoMat E44 PCB milling machine. Aimed at reducing operator training time and improving procedural understanding, the simulation offers an interactive and realistic environment where users can safely engage with machine workflows and start-up sequences. The emphasis is on accurate representation, usability, and maintaining immersion to support intuitive learning. Although formal evaluation is outside the scope of this work, the system is designed to serve as a foundation for cost-effective, scalable training in technical and manufacturing contexts, offering a modern alternative to …

Scalable Real-Time Stream Clustering For Unbounded Text Streams, Nathaniel C. Crossman

Browse all Theses and Dissertations

Social media, AI systems, IoT sensors, and other platforms generate vast amounts of streaming data. Given this vast volume of information, techniques that can reduce and aggregate data into meaningful topics are essential. One such technique is the two-phase stream clustering approach. In the first, online micro-clustering phase, the system forms micro-clusters from the incoming data stream, incrementally merges new items into related existing micro-clusters, and prunes or fades micro-clusters as they become inactive, producing a constantly updating yet compact set of micro-clusters representing potential topics and subtopics of the stream. In the second, offline macro-clustering phase, these micro-clusters are …

Patient Subset Classification Using Encoded Embeddings And Knowledge Graph Retrieval-Augmented Generation, Benjamin A. Holmes

Browse all Theses and Dissertations

The widespread adoption of electronic medical records has created a vast reservoir of clinical data that can be leveraged to better understand how interventions relate to patient outcomes. Much of this information, however, exists as unstructured free-text, posing significant challenges for traditional statistical and machine-learning methods. Solving these challenges would allow the extraction of specific patient subpopulations (clinically relevant cohorts of individuals who share overlapping symptoms, risk factors, or diagnostic criteria), which could be used in precision medicine. Despite this promise, extracting these subpopulations from unstructured medical notes is an ongoing challenge due to the variability of clinical language and …

Generative Adversarial Networks (Gans) For High-Dimensional Biological Data, Harigovind Harikumar

Browse all Theses and Dissertations

This thesis investigates the application of Generative AI models, mainly Generative Adversarial Network (GAN) models to high dimensional and low sample size biological datasets like Motion Sickness, Breast Cancer, Crohn, and Melanoma. We utilized and compared three generative AI frameworks: Vanilla GAN, Wasserstein GAN (WGAN), Locality-Sensitive Hashing GAN (LSH-GAN) and Omics GAN. To address the challenges associated with high-dimensionality and low sample size, which was leading to very poor outputs of biological synthetic samples, we came up with an approach to stop the model when it reaches its saturation level. That is, we printed the loss plots to see where …

Pixmix Attack: Implementation And Evaluation Of A Novel Pixel Injection On Digital Video Port (Dvp) Interface In Embedded Camera Systems With Pcb Hardware Trojan, Sayed Md Tashfi Nowroz

Browse all Theses and Dissertations

Image sensors are at the heart of machine vision systems in robotics, industrial automation, and surveillance systems which ideally operate with minimal human supervision and only occasional maintenance. The image sensors convert visible light into electrical signals which are locally decoded to image on the printed circuit board (PCB) by an ordinary embedded processor System on Chip (SoC). This thesis investigates a critical vulnerability in such systems, targeting the communication protocol at the signal level during runtime. Specifically, it focuses on a novel attack in the Digital Video Port (DVP) protocol, possible to exploit with PCB-based hardware Trojans, to craft …

A Secure Ml-Assisted Framework For Resilient And Efficient Prediction Of Physiotherapy Sequence In Bilateral Carpal Tunnel Syndrome, Pratik Pandurang Kharat

Browse all Theses and Dissertations

Bilateral idiopathic carpal tunnel syndrome (CTS) is a neuromuscular disorder characterized by compression of the median nerve at both wrists, leading to symptoms such as pain, numbness, tingling, and muscle weakness. Unlike unilateral cases, bilateral idiopathic CTS presents distinct therapeutic challenges due to the simultaneous involvement of both hands and the lack of an identifiable underlying cause. This study explores the application of machine learning techniques to predict the optimal sequence of physiotherapeutic interventions Stretching followed by Myofascial Mobilization (S/M) or the reverse (M/S) in female patients with bilateral idiopathic CTS and right hand dominance. Data were drawn from a …

Learning Under Data Scarcity: Reasoning And Negative Distillation For Texts And Graphs, Calvin T. Greenewald

Browse all Theses and Dissertations

Modern machine learning (ML) models rely on large amounts of high-quality labeled data to achieve optimal performance. However, in many real-world domains, such as cyber security, acquiring sufficient labeled data is often infeasible due to cost, privacy concerns, and the rapid evolution of underlying phenomena. This challenge underscores the importance of learning under data scarcity. This thesis addresses this challenge by proposing distinct, modality-specific techniques for text and graph domains, which allow models to generalize effectively with minimal data. For text classification task, we incorporate distilled rationales from large language models and adversarial perturbations into the input space to improve …

Imitation Learning In Robotic Manipulation Using Diffusion Models, Marlon Domingues De Oliveira

Graduate Theses, Dissertations, and Problem Reports (ETD)

This work deals with the problem of teaching robots by demonstration, also known by imitation learning. The main objective of this area of research is to enable the autonomous execution of complex robotic tasks using neural networks trained on expert-generated data, thus allowing the transfer of human knowledge to machines. To this end, this thesis describes an experimental setup especially designed for the study of imitation learning in robotic manipulation tasks and the adaptation and evaluation of a previously published technique to this setup. The experimental setup developed in this work is based on the Franka Emika Research 3 manipulator, …

Predicting Biomechanical Risk Factors For Division - I Women’S Basketball Athletes, Aayushi Shah, Vanaja Agarwal, Dhairya Shah, Harman Jani, Sristi Sharma, Kaya Tolga, Christopher Taber, Mehul Raval

School of Computer Science & Engineering Faculty Publications

Collegiate basketball is characterized by high-impact movements such as jump landings, making athletes more susceptible to injuries. Critical biomechanical factors like knee flexion, lateral trunk flexion, and foot landing asymmetry are strongly associated with injury risk. This study aims to predict six biomechanical risk factors in the landing error scoring system (LESS). The dataset comprises 8600 video frames of counter-movement jumps (CMJs) from 17 NCAA Division I female basketball athletes, recorded from frontal and lateral perspectives and annotated using a customized error annotation algorithm. The study uses the You Only Look Once (YOLOv5nu) model to analyze the basketball athletes’ CMJ …

Implementation Of Quantized Artificial Neural Networks With Spintronic Stochastic Computing, Saadi Sabyasachi Mr.

Theses and Dissertations

Artificial intelligence or machine learning is going through a rapid expansion. It also incurs significant costs for power and device footprints. Various approaches are being explored to design energy and hardware efficient machine learning models. Stochastic computing has been proposed for efficient machine learning implementation. It requires a source of random number generation which poses some practical challenges. So spintronic solutions such as magnetic tunnel junction has been used for random number generation. Again, spintronic random number generation to implement high precision circuit is prone to device-to-device variations. Hence we designed quantized artificial neural network with spintronic stochastic computing which …

Enhancing Ai-Driven Automation For Object Detection And Computer Vision, Nafeeul Alam Walee

College of Graduate Studies: Theses & Dissertations

In recent years, AI-driven automation has revolutionized the field of object detection and computer vision, enabling sophisticated and efficient solutions across various industries. This research explores the latest advances and techniques in improving AI-driven automation for object detection and computer vision applications. We examine state-of-the-art deep learning models and frameworks that have contributed to significant improvements in accuracy and speed and highlight the generative results. The focus is on exploring the real-time processing capabilities that have expanded the applicability of these technologies in real-world scenarios. Furthermore, we investigate image integration and video data to improve precision detection and contextual understanding. …

Cognitive Map Generation For Vision And Language Navigation, Alexander Sandoval Mesa

Dissertations and Theses

Visual-Language Navigation (VLN) presents significant challenges for autonomous agents, such as robots and virtual assistants, particularly in complex, dynamic environments where the seamless integration of visual perception and natural language understanding is critical. Traditional VLN systems often struggle with effectively aligning language instructions and visual scene understanding, limiting their adaptability and navigation efficiency.

This thesis proposes a novel Cognitive Map-based framework that addresses these challenges by transforming natural language navigation instructions into structured graph representations. The Cognitive Map consists of nodes representing waypoints, landmarks, decision points, and edges encoding spatial relationships and navigational actions. These maps are generated using Large …

Machine Learning-Driven Optimization For Utility-Scale Quantum Optimization, Bao Tran

Theses and Dissertations

Hard combinatorial optimization problems, often mapped to Ising models, promise potential solutions with quantum advantage but are constrained by limited qubit counts in near-term devices. We present an innovative quantum-inspired framework that dynamically compresses large Ising models to fit available quantum hardware of different sizes. Thus, we aim to bridge the gap between large-scale optimization and current hardware capabilities. Our method leverages a physics-inspired GNN architecture to capture complex interactions in Ising models and accurately predict alignments among neighboring spins (aka qubits) at ground states. By progressively merging such aligned spins, we can reduce the model size while preserving the …

Swarming Segregation: Leveraging Swarm Intelligence And Regionalization As Instruments For School District Desegregation, Jeffrey Wooten

Theses and Dissertations

Even after Brown led to the South briefly having the most diverse schools in the nation, schools throughout the Northeast have remained the most segregated in the nation for decades. While federal jurisprudence has made compelling desegregation pursuant to the Equal Protection Clause more challenging, New Jersey has a particularly favorable landscape to address severe segregation. With a highly diverse, densely populated public enrollment, favorable state constitutional precedent, and a history of successfully compelling desegregation, New Jersey is fertile ground exploring regional desegregation. Scholars, judges, and even plaintiffs in ongoing litigation (Latino Action Network v. N.J.) have called for New …

Quantum-Resilient Architectures For Enterprise And Cloud Information Systems, Implications Of Quantum Computing For Enterprise Cybersecurity And Data Integrity, Manikantha Varaprasad Inakollu

Computer Science and Engineering Faculty Publications

The emergence of quantum computing presents unprecedented challenges to contemporary enterprise cybersecurity frameworks. Current cryptographic systems that protect sensitive data and secure communications will become vulnerable to quantum attacks within the next decade. This research examines the implications of quantum computing advancement for enterprise and cloud information systems, proposing quantum-resilient architectural frameworks that can withstand both classical and quantum threats. We analyze the timeline of quantum computing development, assess vulnerabilities in existing enterprise security infrastructures, and evaluate post-quantum cryptographic approaches suitable for organizational implementation. Through comparative analysis of quantum-resistant algorithms and architectural patterns, this study demonstrates that enterprises must begin …

Blockchain-Enabled Trust Frameworks For Enterprise Information Systems, Establishing Verifiable Trust In Distributed Organizational Environments, Manikantha Varaprasad Inakollu

Computer Science and Engineering Faculty Publications

Enterprise information systems increasingly operate in distributed environments where traditional trust mechanisms based on centralized authority prove insufficient. This research develops a comprehensive blockchain-enabled trust framework that establishes verifiable, decentralized trust mechanisms for enterprise systems operating across organizational boundaries. The study addresses critical gaps in current enterprise architectures where trust depends on centralized intermediaries, creating single points of failure and limiting inter-organizational collaboration. Through examination of existing trust models and blockchain capabilities, we propose an integrated framework that combines cryptographic verification, distributed consensus, and smart contract automation to establish trust without centralized control. Our framework enables organizations to verify data …

Detection Of Data Leakage And Disruption Of Covert Timing Channel In Secure Drone Communication Using Machine And Deep Learning, Jonathan Walatkiewicz

Master's Theses and Doctoral Dissertations

The utilization of recreational drones has experienced a substantial increase in both the United States and globally. However, it is noteworthy that most drones, classified as Internet of Things devices, are produced with a limited security lifecycle. This study's findings are of paramount importance, as traditional computing exploits can be applied to drones, designating them as high- value targets. This study examines the detectability and disruptability of covert timing channel traffic in secure drones. The investigation aims to ascertain the effects of multiple interarrival times, distances ranging from 1 to 330 feet, various detection algorithms, and stream sizes between 32-bit …

Motion Planning For A Flexible Modular Raft Robot, Chun-Yi She

Dartmouth College Master’s Theses

This thesis presents a hierarchical motion planning framework for SoftRafts, a modular and deformable aquatic robot capable of performing locomotion and manipulation tasks on water surfaces. SoftRafts consist of soft and rigid components that enable structural reconfiguration, offering adaptability in unstructured aquatic environments.

To address the complexity of planning in high-dimensional, deformable systems, the proposed method uses a bounding-shape abstraction, specifically, enclosing circles and rectangular bounding boxes to simplify motion planning. These enclosures abstract the robot's overall shape, reducing the high-dimensional planning problem into a lower-dimensional problem. A global planner uses a probabilistic roadmap (PRM) to compute a collision-free path …