Software Engineering Commons^™

Evolving Enemy Behavior In Video Games, Hermie H. Adams Iii

Honors Theses

The video game I developed for my senior project lacked complex and engaging enemy artificial intelligence. The standard implementations of AI systems such as finite state machines and behavior trees felt like side-steps rather than innovative solutions. Upon seeing the 'magic' of machine learning in perfecting games such as Snake, Super Mario, and Flappy Bird, I was inspired to seek my answer in the field of evolutionary computation. However, my challenge differed in that the problem space would be defined by dynamic player strategies, making it not well-defined or static. As such, my evaluations are based on enemies exhibiting emergent …

Improving The Reproducibility Of Deep Learning Software: An Initial Investigation Through A Case Study Analysis, Nikita Ravi, Abhinav Goel, James C. Davis, George K. Thiruvathukal

Computer Science: Faculty Publications and Other Works

The field of deep learning has witnessed significant breakthroughs, spanning various applications, and fundamentally transforming current software capabilities. However, alongside these advancements, there have been increasing concerns about reproducing the results of these deep learning methods. This is significant because reproducibility is the foundation of reliability and validity in software development, particularly in the rapidly evolving domain of deep learning. The difficulty of reproducibility may arise due to several reasons, including having differences from the original execution environment, incompatible software libraries, proprietary data and source code, lack of transparency, and the stochastic nature in some software. A study conducted by …

Hybridize Functions: A Tool For Automatically Refactoring Imperative Deep Learning Programs To Graph Execution, Raffi Khatchadourian, Tatiana Castro Vélez, Mehdi Bagherzadeh, Nan Jia, Anita Raja

Publications and Research

Efficiency is essential to support responsiveness w.r.t. ever-growing datasets, especially for Deep Learning (DL) systems. DL frameworks have traditionally embraced deferred execution-style DL code—supporting symbolic, graph-based Deep Neural Network (DNN) computation. While scalable, such development is error-prone, non-intuitive, and difficult to debug. Consequently, more natural, imperative DL frameworks encouraging eager execution have emerged but at the expense of run-time performance. Though hybrid approaches aim for the “best of both worlds,” using them effectively requires subtle considerations to make code amenable to safe, accurate, and efficient graph execution—avoiding performance bottlenecks and semantically inequivalent results. We discuss the engineering aspects of a …

Hybridize Functions: A Tool For Automatically Refactoring Imperative Deep Learning Programs To Graph Execution, Raffi Khatchadourian, Tatiana Castro Vélez, Mehdi Bagherzadeh, Nan Jia, Anita Raja

Publications and Research

Efficiency is essential to support responsiveness w.r.t. ever-growing datasets, especially for Deep Learning (DL) systems. DL frameworks have traditionally embraced deferred execution-style DL code—supporting symbolic, graph-based Deep Neural Network (DNN) computation. While scalable, such development is error-prone, non-intuitive, and difficult to debug. Consequently, more natural, imperative DL frameworks encouraging eager execution have emerged but at the expense of run-time performance. Though hybrid approaches aim for the “best of both worlds,” using them effectively requires subtle considerations to make code amenable to safe, accurate, and efficient graph execution—avoiding performance bottlenecks and semantically inequivalent results. We discuss the engineering aspects of a …

Hybridize Functions: A Tool For Automatically Refactoring Imperative Deep Learning Programs To Graph Execution, Raffi Khatchadourian, Tatiana Castro Vélez, Mehdi Bagherzadeh, Nan Jia, Anita Raja

Publications and Research

Efficiency is essential to support responsiveness w.r.t. ever-growing datasets, especially for Deep Learning (DL) systems. DL frameworks have traditionally embraced deferred execution-style DL code—supporting symbolic, graph-based Deep Neural Network (DNN) computation. While scalable, such development is error-prone, non-intuitive, and difficult to debug. Consequently, more natural, imperative DL frameworks encouraging eager execution have emerged but at the expense of run-time performance. Though hybrid approaches aim for the "best of both worlds," using them effectively requires subtle considerations to make code amenable to safe, accurate, and efficient graph execution—avoiding performance bottlenecks and semantically inequivalent results. We discuss the engineering aspects of a …

Computer Vision In Soccer: Yolov11 Analytics Engine For Quantifying Game Strategy, Connor S. Maurer

Data Science Undergraduate Honors Theses

Single-shot object detection capabilities significantly reduce computational overhead for real-time computer vision in sports analytics at 60 FPS. YOLO11’s lightweight CNN gives promising accuracy while meeting the low-latency demand of dynamic soccer matches. As data-driven approaches take over the sport of soccer, efficient player tracking systems become critical for informing coach’s strategies. I prototype the ETL (Extract, Transform, Load) process of data collected from a single- shot detection program and evaluate its viability for estimating player fatigue. YOLO11 detects players, the ball, and other characteristics, with the output transformed by homography to estimate the positions in the real world. These …

Ai And Prompt Engineering For Library Discovery Services, James Day

Publications

We have seen the rise of generative artificial intelligence in the form of Large Language Models (LLMs) to provide answers to users’ queries. Services such as ChatGPT, Copilot, and Gemini have quickly become accepted and adopted in the research process. Now library vendors are adding artificial intelligence (AI) to their discovery services to allow for natural language queries to produce generative results. However, the AI model used for discovery services differs from normal LLMs in a significant way that has several positive benefits, but it affects how prompts are written. Library discovery services use a model called Retrieval- Augmented Generation …

Applying Software Engineering Black-Box Methods For Testing Machine Learning Models, Timothy Elvira

Doctoral Dissertations and Master's Theses

This dissertation proposes researching an approach to incorporate and align Software black-box testing methods into Machine Learning (ML) applications, specifically in the context of computer vision models. Typically, testing methods within Software Engineering (SE) encompass a range of test types that assess levels of a software system, such as Unit, Integration, Functional, and System testing [1]. The testing spectrum offers two perspectives on the system: black-box, where the system’s code is hidden, and white-box, where the system's code is exposed for testing. Software Quality pairs testing with requirements, in a many-to-one relationship, to ensure proper validation of the software system. …

Bigcodebench: Benchmarking Code Generation With Diverse Function Calls And Complex Instructions, T.Y. Zhuo, M.C. Vu, J. Chim, ..., David Lo

Research Collection School Of Computing and Information Systems

Task automation has been greatly empowered by the recent advances in Large Language Models (LLMs) via Python code, where the tasks ranging from software engineering development to general-purpose reasoning. While current benchmarks have shown that LLMs can solve tasks using programs like human developers, the majority of their evaluations are limited to short and self-contained algorithmic tasks or standalone function calls. Solving challenging and practical tasks requires the capability of utilizing diverse function calls as tools to efficiently implement functionalities like data analysis and web development. In addition, using multiple tools to solve a task needs compositional reasoning by accurately …

Adapting Knowledge Prompt Tuning For Enhanced Automated Program Repair, Xuemeng Cai, Lingxiao Jiang

Research Collection School Of Computing and Information Systems

Automated Program Repair (APR) aims to enhance software reliability by automatically generating bug-fixing patches. Recent work has improved the state-of-the-art of APR by fine-tuning pre-trained large language models (LLMs), such as CodeT5, for APR. However, the effectiveness of fine-tuning be-comes weakened in data scarcity scenarios, and data scarcity can be a common issue in practice, limiting fine-tuning performance. To alleviate this limitation, this paper adapts prompt tuning for enhanced APR and conducts a comprehensive study to evaluate its effectiveness in data scarcity scenarios, using three LLMs of different sizes and six diverse datasets across four programming languages. Prompt tuning rewrites …

Adaptive Deviation Learning For Visual Anomaly Detection With Data Contamination, Aanindya Sundar Das, Guansong Pang, Monowar Bhuyan

Research Collection School Of Computing and Information Systems

Visual anomaly detection targets to detect images that notably differ from normal pattern, and it has found extensive application in identifying defective parts within the manufacturing industry. These anomaly detection paradigms predominantly focus on training detection models using only clean, unlabeled normal samples, assuming an absence of contamination; a condition often unmet in real-world scenarios. The performance of these methods significantly depends on the quality of the data and usually decreases when exposed to noise. We introduce a systematic adaptive method that employs deviation learning to compute anomaly scores end-to-end while addressing data contamination by assigning relative importance to the …

Revisiting Sentiment Analysis For Software Engineering In The Era Of Large Language Models, Ting Zhang, Ivana Clairine Irsan, Thung Ferdian, David Lo

Research Collection School Of Computing and Information Systems

Software development involves collaborative interactions where stakeholders express opinions across various platforms. Recognizing the sentiments conveyed in these interactions is crucial for the effective development and ongoing maintenance of software systems. For software products, analyzing the sentiment of user feedback, e.g., reviews, comments, and forum posts can provide valuable insights into user satisfaction and areas for improvement. This can guide the development of future updates and features. However, accurately identifying sentiments in software engineering datasets remains challenging.This study investigates bigger large language models (bLLMs) in addressing the labeled data shortage that hampers fine-tuned smaller large language models (sLLMs) in software …

Evaluating Software Development Agents: Patch Patterns, Code Quality, And Issue Complexity In Real-World Github Scenarios, Zhi Chen, Lingxiao Jiang

Research Collection School Of Computing and Information Systems

In recent years, AI-based software engineering has progressed from pre-trained models to advanced agentic workflows, with Software Development Agents representing the next major leap. These agents, capable of reasoning, planning, and interacting with external environments, offer promising solutions to complex software engineering tasks. However, while much research has evaluated code generated by large language models (LLMs), comprehensive studies on agent-generated patches, particularly in real-world settings, are lacking. This study addresses that gap by evaluating 4,892 patches from 10 top-ranked agents on 500 real-world GitHub issues from SWE-Bench Verified, focusing on their impact on code quality. Our analysis shows no single …

Bridging Expert Knowledge With Deep Learning Techniques For Just-In-Time Defect Prediction, Xin Zhou, Donggyun Han, David Lo

Research Collection School Of Computing and Information Systems

Just-In-Time (JIT) defect prediction aims to automatically predict whether a commit is defective or not, and has been widely studied in recent years. In general, most studies can be classified into two categories: 1) simple models using traditional machine learning classifiers with hand-crafted features, and 2) complex models using deep learning techniques to automatically extract features from commit contents. Hand-crafted features used by simple models are based on expert knowledge but may not fully represent the semantic meaning of the commits. On the other hand, deep learning-based features used by complex models represent the semantic meaning of commits but may …

Towards Resource-Efficient Reactive And Proactive Auto-Scaling For Microservice Architectures, Hussain Ahmad, Christoph Treude, Markus Wagner, Claudia Szabo

Research Collection School Of Computing and Information Systems

Microservice architectures have become increasingly popular in both academia and industry, providing enhanced agility, elasticity, and maintainability in software development and deployment. To simplify scaling operations in microservice architectures, container orchestration platforms such as Kubernetes feature Horizontal Pod Auto-scalers (HPAs) designed to adjust the resources of microservices to accommodate fluctuating workloads. However, existing HPAs are not suitable for resource-constrained environments, as they make scaling decisions based on the individual resource capacities of microservices, leading to service unavailability, resource mismanagement, and financial losses. Furthermore, the inherent delay in initializing and terminating microservice pods hinders HPAs from timely responding to workload fluctuations, …

Heartdj - Music Recommendation And Generation Through Biofeedback From Heart Rate Variability, Egemen Şahin

Dartmouth College Master’s Theses

This study investigates the integration of real-time physiological data with AI-generated music to enhance emotional well-being, stress regulation, and focus, using Heart Rate Variability (HRV) as a biomarker of autonomic function. Conducted in two phases—Stable Audio Open (SAO) and Suno (SUNO)—the research evaluates biofeedback-driven music interventions across varying daily music-listening habits.

In the SAO phase, short AI-generated instrumental tracks were compared with Spotify recommendations and guided meditation. Modest HRV improvements were observed in biofeedback conditions, but participants noted emotional limitations, citing short track lengths and abrupt transitions.

The SUNO phase addressed these limitations with longer, more complex AI-generated compositions combined …

Rotary Operations Management & Automation Platform (Romap): Modernizing Attendance Tracking And Data Submission For Rotary Clubs, Zaineel Mithani

2025 Fall Honors Capstones Projects - Archive

As Technical Lead of the Rotary Operations Management & Automation Platform (ROMAP), my Honors contribution focused on developing a Bluetooth Low Energy proximity-based attendance system enabling automatic, hands-free member check-ins. I researched and selected beacon hardware, designed RSSI-based distance calculation algorithms, and implemented platform-specific background processing for iOS and Android, achieving 97% detection accuracy. Beyond this Honors component, I architected the complete backend infrastructure including a Node.js API with 20+ endpoints, PostgreSQL database with Prisma ORM, and JWT authentication. I also developed a novel GPT-4 Vision automation system that intelligently populates web forms through computer vision, achieving 95% success rate …

Towards Smart Farming: Image-Based Crop Health Assessment And Disease Diagnosis Using Deep Learning Techniques, Kristina Botova

Master's Theses or Doctor of Nursing Practice

Accurate crop monitoring is essential for optimizing agricultural productivity and ensuring food security. This study presents a comprehensive deep learning framework for image crop type recognition, health status prediction, and disease detection using multiple Convolutional Neural Network (CNN) models. The proposed approach uses open-source datasets consisting of five crop types (apple, corn, grape, potato, tomato), varying health conditions, and common diseases. By deploying specialized CNN architecture focused on each task, the system achieves a high accuracy of 99.25% in classifying crop types, identifying health status, and detecting specific diseases. Compared to a single CNN model, the use of the proposed …

Csc36000 - Modern Distributed Computing Assignment 2, Saptarashmi Bandyopadhyay

Open Educational Resources

This assignment is designed to help the student identify and mitigate common errors in Distributed Computing such as race conditions and reaching consensus, as well as reflecting on how Distributed Computing concepts apply to their class project.

Cmc Thesis Chatbot, Luis Gomez

CMC Senior Theses

This GitHub repo is a senior thesis for Claremont McKenna College; it is a thesis about theses. The project is an interactive RAG-based chatbot that helps students, researchers, and faculty explore Claremont McKenna College senior theses. The goal was to create a domain-specific chatbot to show that it is possible to combat the limitations of AI, including hallucinations, outdated data, and lack of domain expertise. The website link is:

CMCThesisChatbot

Understanding Physiological Responses For Intelligent Posture Detection Using Wearable Technology, Chaitanya Vardhini Anumula, Tanvi Banerjee, Anuradha Oak

Celebration of Undergraduate & Graduate Research, Scholarship, and Creative Activities Materials

This study investigates the physiological impact of Iyengar yoga at the pose-level using EmbracePlus wearable smartwatch, for data recording and personalized yoga pose detection for tracking.

Temporal Machine Learning For Predicting Accidents And Violations In The Mining Industry, Nathan T. Kelley

Theses and Dissertations--Mining Engineering

This thesis examines the predictive capability of a temporal machine learning model for forecasting future accidents and violations at individual mines, based on historical data. Mine accidents were categorized by accident classification and violations were categorized by the Part Section. The primary datasets utilized were the mine safety and health administration’s (MSHA’s) Accident Injuries and Violations datasets. The available datasets were cleaned and organized by mine type and commodity, then divided into separate subsets for training, validating, and testing. Different models, cutoff metrics, learning rates, number of hidden layers, data processing methods, data processing divisions, number of points observed …

Measuring Model Alignment For Code Clone Detection Using Causal Interpretation, Shamsa Abid, Xuemeng Cai, Lingxiao Jiang

Research Collection School Of Computing and Information Systems

Deep Neural Network-based models have demonstrated high accuracy for semantic code clone detection. However, the lack of generalization poses a threat to the trustworthiness and reliability of these models. Furthermore, the black-box nature of these models makes interpreting the model’s decisions very challenging. Currently, there is only a limited understanding of the semantic code clone detection behavior of existing models. There is a lack of transparency in understanding how a model identifies semantic code clones and the exact code components influencing its prediction. In this paper, we introduce the use of a causal interpretation framework based on the Neyman-Rubin causal …

Automatic Scoring Cornhole Board System, Eric Diffendal, Jonah Harsh, Connor Lengel, Brett Sukie

Williams Honors College, Honors Research Projects

The objective is to create a self-scoring cornhole board that can detect and calculate each team's score based on the bags thrown each round and to be created at a low cost/eventually being sold at the current cost of a normal board. When playing cornhole, the game is simple: throw a bag on the board; however, the scores are variable (deduct and add) across each round. The most common issue when playing cornhole is miscalculations of the scores and forgetting the correct scores. Thus, this invention will make gameplay easy for all to play.

Skinrisk Ai, Spencer Simms

Williams Honors College, Honors Research Projects

SkinRisk AI is an exploration of the opportunities for implementing machine learning (ML) and artificial intelligence (AI) in the medical technology field, specifically in the early detection of skin cancer. This project presents the design, development, and evaluation of a mobile application that allows users to capture images of skin lesions and receive a machine learning assisted risk assessment. The system combines a convolutional neural network (CNN) for image analysis with an intuitive mobile app built using Flutter, FastAPI, and Supabase to deliver real time screening.

Motivated by the rising skin cancer rates and importance of early detection, SkinRisk AI …

Large Scale Machine Learning Over Knowledge Graphs, Bedirhan Gergin

Electronic Theses & Dissertations (2024 - present)

Knowledge graphs (KGs) have become popular across various fields, providing convenient access to web-based knowledge while storing and formalizing domain-specific information. By analyzing KGs, patterns, connections, and dependencies can be identified across different data sources, enabling the inference of new knowledge from given facts. As the use of KGs expands, the size of modern KGs has grown significantly, making them impossible to process within the main memory of a single computer. Distributed computing offers a viable solution to this challenge by leveraging the combined capabilities of multiple servers within a cluster. This thesis explores how distributed computing can be effectively …

Openmuse: Integrating Open-Source Models Into Music Creation Workflows, Tyler K. Vergho

Dartmouth College Master’s Theses

This master's thesis introduces OpenMUSE (Open Multimodal Unified Sound Engine), a platform that demonstrates the potential of open-source AI music generation by integrating state-of-the-art deep learning models into a unified system. By unifying ten different open-source models, including MusicGen, AudioLDM2, and custom-trained text-to-symbolic music generation models, OpenMUSE aims to create a user-friendly interface that empowers artists to produce complex, adaptive musical compositions. The system enhances accessibility by providing a simple web interface and natural language controls, while improving controllability through features like melody conditioning and semantic audio editing. Specifically, OpenMUSE offers a digital audio workstation (DAW)-inspired interface that lowers the …

Neural And Computational Approach To Understanding Environmental Modulation Of Behavioral Identity In Zebrafish, John W. Hageter

Graduate Theses, Dissertations, and Problem Reports (ETD)

Organisms rely on behavior for survival. Animals engage in behaviors that allow for feeding, mating, exploring and navigating their environment among others. Necessary for these behaviors to develop are the environmental factors and underlying circuitry which make behavior possible. Specifically, how the environment guides underlying neural circuitry to develop unique facets or phenotypes of a larger behavior are key to understanding why unique behaviors exist. In this thesis, I build foundational evidence for determining these mechanisms through the use of the zebrafish local search behavior. This is a behavior that zebrafish employ following the loss of environmental illumination where they …

Graph Neural Networks Powered Scientific Paper Recommendation, Junhao Shen

Computer Science and Engineering Theses and Dissertations

Scientific paper recommendation systems aim to help researchers discover relevant papers amidst the vast and ever-growing body of literature. With the exponential yearly increase in scientific publications, the demand for effective paper recommendation solutions has become both critical and increasingly challenging. In recent years, deep learning techniques have revolutionized recommender systems, and scientific paper recommendations have naturally integrated these advancements. In this dissertation, we address these challenges through three progressive contributions.

First, we enhance traditional content-based methods using Graph Neural Networks (GNNs) by introducing a Graph Convolutional Network-strengthened Topic Modeling (GCN-TM) approach. This method improves upon conventional topic modeling techniques …

Computational Representation, Analysis And Verification Of Requirements In Engineering Design And Systems Engineering, Chandan Kumar Sahu

All Dissertations

Systems are developed to satisfy a set of requirements derived from stakeholders’ needs, defining the problem space for which the system is created as a feasible solution. The system design process begins with eliciting these requirements and concludes with validating whether the created system meets them. Requirements engineering (RE) encompasses elicitation, representation, analysis, documentation, verification, and validation. However, challenges in RE, such as imprecision in natural language (NL), proprietary restrictions, and a lack of standardized quality metrics, hinder the creation of well-formed and comprehensive requirements. These challenges complicate formalization and analysis of requirements.

This dissertation addresses these challenges by proposing …