Digital Commons Network^™

Opening The Black Box With Regal: A Novel Explainable Ai Approach To Uncover Key Predictors In Search And Rescue Success, Brandon Hyunjun Kim

Master's Theses

The outcome of a search and rescue (SAR) operation is influenced by a complex, non-linear interplay among numerous factors, including geographic context, subject-specific characteristics, and environmental conditions. The high dimensionality and intricate dependencies among these variables pose significant challenges to traditional exploratory modeling approaches, limiting their ability to uncover meaningful patterns and relationships associated with mission success. This study introduces Rules Based Explanations for Generated neighborhoods Around Localized cases (REGAL), a novel adaptation of the Local Interpretable Model-agnostic Explanations (LIME) framework to explain deep multimodal neural networks and what key features it assesses to determine search and rescue success. REGAL …

Efficient Gan-Based Adversarial Example Generation Against Ml-Based Network Intrusion Detection Systems, Darren D. Hartono

Master's Theses

In the realm of network security, Network Intrusion Detection Systems (NIDS) are essential for identifying and mitigating malicious activities targeting networked devices. Traditionally, these systems have relied on signature-based and anomaly-based detection techniques. However, the increasing complexity and adapt- ability of cyber threats have driven the adoption of Machine Learning (ML) ap- proaches in modern NIDS, significantly improving their ability to detect a wider range of attack vectors. Despite these advancements, ML-based NIDS remain vulnerable to adversarial examples—deliberately crafted inputs designed to mislead models and trigger incorrect classifications. Originally identified in the field of computer vision, adversarial examples now pose …

Efficient Blending Of Large Language Models, Sandeep Chatterjee

Master’s Dissertations

Due to the limited capabilities of single Large Language Models (LLMs), multiple LLMs can be employed in tandem for better reliability of answers. Blending refers to combining the strengths of various LLMs to make use of their complementary capabilities for generating high-quality responses. It is a non-trivial problem, and the task becomes even more difficult when aiming for minimal latency and supervising the blending components. The standard framework, LLM- Blender, approaches this in three stages: response generation, candidate selection via ranking, and response fusion through summarization. However, this pipeline faces two critical limita- tions—high latency due to repeated ranking steps, …

Computerized Diagnostic Decision Support Systems-Isabel Pro Versus Chatgpt-4 Part Ii, Joe M Bridges, Xiaoqian Jiang, Michael Ige, Oluwatoniloba Toyobo

Faculty, Staff and Student Publications

Objective: Does a Tree-of-Thought prompt and reconsideration of Isabel Pro's differential improve ChatGPT-4's accuracy; does increasing expert panel size improve ChatGPT-4's accuracy; does ChatGPT-4 produce consistent outputs in sequential requests; what is the frequency of fabricated references?

Materials and methods: Isabel Pro, a computerized diagnostic decision support system, and ChatGPT-4, a large language model. Using 201 cases from the New England Journal of Medicine, each system produced a differential diagnosis ranked by likelihood. Statistics were Mean Reciprocal Rank, Recall at Rank, Average Rank, Number of Correct Diagnoses, and Rank Improvement. For reproducibility, the study compared the initial expert panel run …

Verify All Traffic: Towards Zero-Trust In-Network Intrusion Detection Against Multipath Routing, Ziming Zhao, Zhaoxuan Li, Xiaofei Xie, Zhipeng Liu, Tingting Li, Jiongchi Yu, Fan Zhang, Binbin Chen

Research Collection School Of Computing and Information Systems

With the popularity of encryption protocols, machine learning (ML)-based traffic analysis technologies have attracted widespread attention. To adapt to modern high-speed bandwidth, recent research is dedicated to advancing zero-trust intrusion detection by offloading feature extraction and model inference into the network dataplane. Especially, with the rise of programmable switches, achieving line-speed ML inference becomes promising. However, existing research only considers a single switch node as a relay to conduct evaluation. This is far from real-world deployments involving multiple switches (given that zero-trust security assumes that threats can originate from anywhere, including within the network), particularly the multipath routing phenomenon that …

Digital Resurrection Of Thonis-Heracleion: Technological Advances In Underwater Archaeology And A Speculative Ai-Driven Reconstruction Methodology, James Hutson, Passent Chahine

Faculty Scholarship

This article synthesizes past archaeological research on the submerged Egyptian city of Thonis-Heracleion, critically reviewing excavations and technological interventions deployed since its rediscovery by Franck Goddio and the IEASM team. Situated approximately 10 meters beneath Aboukir Bay near Alexandria, the city represents a significant nexus of Greek and Egyptian cultural heritage, vividly documented in classical sources such as Herodotus and Strabo. Prior excavations have recovered temple complexes, colossal statues, ritual artifacts, and an extensive array of ancient shipwrecks, mapping only a fraction of the extensive site. These investigations utilized pioneering geophysical methods, including multibeam sonar, side-scan sonar, and photogrammetry, establishing …

Coda: A Digital System For Generation Of Piano Practice Exercises From Symbolic Music Notation, Annie Tang

Computer Science Senior Theses

Effective practice remains one of the greatest challenges in music education, yet cur- rent digital music tools primarily support only passage engagement or surface-level feedback, failing to provide proactive guidance for overcoming technical challenges within piano repertoire. This thesis presents Coda, a digital system that generates custom piano practice exercises from symbolic music notation based on established piano pedagogy principles that have historically been taught orally.

Unlike existing tools that only provide a viewable symbolic music notation display or only an interface to take notes and record a practice session, Coda uses rule- based algorithmic transformations rooted in pedagogical logic …

Alibaba: Building Advanced Intellectual Property Governance For E-Commerce Marketplaces, Liang Chen, Sin Mei Cheah, Can Huang, Guoqiao Liu

Asian Management Insights

How the global e-commerce powerhouse harnessed artificial intelligence (AI) to balance innovation and intellectual property (IP) rights protection.

Topodino: Self-Supervised Topological Representation Learning For Neuronal Morphologies, Yasser Binbisher

Master's Theses

Neuronal cell types are categorized by transcriptomic identity, yet their morphological heterogeneity defies this classification. In response, researchers have adopted unsupervised graph representation learning as a tool to reveal morphological variation within single-class transcriptomic types. However, the complex geometry of neuronal morphology—especially long axons and dense dendrites—challenges graph neural networks, which struggle with message propagation across extended structures. To mitigate this, current approaches enforce sub-sampling on neuronal graphs and omit axons entirely, sacrificing critical biological features for computational efficiency. To overcome this trade-off, this thesis introduces TopoDINO, a self-supervised, topology-aware representation learning model designed to preserve the full hierarchical organization …

Exploiting Compiler-Introduced Vulnerabilities In C: A Cross-Compiler And Cross-Architecture Analysis Of Undefined Behavior, Erik Mccutchen

Master's Theses

Compilers are a critical component in generating secure software across engineering disciplines. However, languages like C that permit undefined behavior introduce a fundamental tension between the compiler’s interpretation of undefined behavior and the security of the generated code. This tension can result in security vulnerabilities that, from the programmer's perspective, are ``created'' by the compiler. The widespread use of these languages, combined with the complexity of modern optimizations and limited developer visibility into compiler behavior, makes these vulnerabilities both pervasive and difficult to detect.

Building on prior work, this thesis refines a dataset of C code snippets that exhibit Compiler-Introduced …

Enhancing Energy Consumption Forecasting For Electric Vehicle Charging Stations With Time Series Dense Encoder (Tide), Amril Nazir, Abdul Khalique Shaikh, Aftab Ahmed Khan, Abdul Salam Shah, Nadia Khalique

All Works

The increasing adoption of electric vehicles has led to the installation of charging stations in various locations in major cities worldwide. This study focuses on energy consumption forecasting for Boulder, Nevada, United States electric vehicle charging stations. Efficient management of energy resources at these charging points is crucial for optimizing resource utilization and reducing charging time. While existing literature has focused on energy consumption prediction in smart homes and grids, the significance of electric charging points in smart cities must be considered. The transformers have handled time series forecasting better with larger datasets like the Temporal Fusion Transformer and the …

An Exposition Of "Probabilistic Polynomials And Hamming Nearest Neighbors", Vivek Srirama

University Honors Theses

This paper is an exposition of the paper Probabilistic Polynomials and Hamming Nearest Neighbors by Josh Alman and Ryan Williams. It presents the findings of this paper in a more accessible format for Computer Science students earlier in their career who may not be as familiar with Computational Theory and its concepts as their PhD counterparts are. The paper assumes that the reader has a basic understanding of Algorithms and Complexity, typically obtained in an introductory level Algorithms course.

The paper by Alman and Williams analyzes a specific problem known as the Hamming Nearest Neighbor problem. All known solutions for …

Contract Quality Feature Extraction Using Llm, Aaron C. Washington

Theses and Dissertations

This study explored the potential insights generated from linguistic complexity measurements and large language model (LLM) based assessments on the quality of contract documents. By combining structured True/False prompts with log-probability analysis and ambiguity scoring, the study introduced novel contract-quality assessment methods. Results support a feature-driven approach to contract evaluation, one that offers automated, scalable insights for triaging risk and improving drafting practices. These assessment methods contribute to the growing field of legal natural language processing by offering modular tools for effective contract analysis.

Glacial Lakes Segmentation Using Multispectral Remote Sensing Data And Deep Learning Models, Debankan Das

Master’s Dissertations

The identification and delineation of glacial lakes through segmentation is crucial for tracking glacial changes and evaluating potential dangers from sudden flood events (GLOFs). These floods can severely impact populated areas and man-made structures downstream. Recent advances in high-quality satellite imagery have sparked increased attention toward using advanced machine learning methods, particularly deep learning, to enable precise and automated glacial lake detection. In this study, we explore the effectiveness of deep learning-based pointwise semantic segmentation for glacial lake mapping using multisource remote sensing imagery, including both optical and synthetic aperture radar (SAR) data. We experiment with a novel stack combination …

Modeling And Verification Of Sigma Delta Neural Networks, Sirshendu Das

Master’s Dissertations

In the context of modern day embedded safety-critical systems and low-resource edge devices in particular, Sigma-Delta Neural Networks (SDNNs) offer a promising alternative to traditional Artificial Neural Networks (ANNs) by leveraging eventdriven, sparse computations inspired by biological neural processing. This energyefficient paradigm makes SDNNs well-suited for neuromorphic hardware and realtime applications, particularly in scenarios with temporal redundancy, such as video processing. However, as neural networks become integral to safety-critical systems, ensuring their robustness against adversarial perturbations is an absolute necessity. In this work, we propose an end-to-end framework for formal modeling and verification of SDNNs using Satisfiability Modulo Theory (SMT). …

Addressing Class Imbalance Problems To Improve Animal Detection Through Aerial Image Data, Suryang Koushal

Master’s Dissertations

Monitoring animal populations in wildlife reserves is essential for conservation, especially for endangered species, but manual censuses are costly, risky, and logistically challenging due to vast, inaccessible terrains. Unmanned Aerial Vehicles (UAVs) with digital cameras provide a safer, scalable solution for collecting aerial imagery to estimate animal populations. However, semi-automated processing of these images faces significant challenges due to class imbalance in datasets, including foreground-background disparities, where background terrain dominates over sparse animal instances, and inter-class imbalances from uneven species representation and varied visual appearances (e.g., species, sizes, fur patterns) against diverse backgrounds like deserts or forests. These imbalances hinder …

Enhancing Expressive Power Of Graph Neural Networks Using Geometric Transformations, Suranjan Dey

Master’s Dissertations

Graph Neural Networks (GNNs) are highly effective in many real-world tasks, such as molecular property prediction, modeling protein structures, analyzing user-item relationships, and making link predictions. What sets them apart is their ability to learn meaningful representations by capturing not just the features of individual nodes, but also the overall structure of the graph they belong to. This expressive strength allows GNNs to model complex relationships more accurately. In this work, we take a step further by introducing geometric transformations aimed at improving how GNNs handle spatial information. In particular, we focus on angular aggregation methods that maintain rotational consistency, …

Application Of Deep Learning In Analysis Of Stellar Spectra, Piyush Yayati

Master’s Dissertations

In recent years, the analysis of high-resolution stellar spectra has become increasingly important for estimating key stellar parameters such as effective temperature (Teff ), surface gravity (log g), metallicity ([M/H]), and rotational velocity (v sin i). Traditional methods often rely on manual calibration or spectrum synthesis, which can be time-consuming and error-prone, especially for M dwarfs whose spectra are dense with molecular features. In this study, we investigate the use of convolutional neural networks (CNNs) to automate the estimation of stellar parameters using synthetic and observed data.We adopt a StarNet-like CNN architecture trained on synthetic spectra generated from the PHOENIX-ACES …

Interactive Generative Modeling: A Pathway For Improved Simulation And Decision Making, Changyu Chen

Dissertations and Theses Collection (Open Access)

This dissertation presents Interactive Generative Modeling (IGM), a unified perspective that integrates interactive paradigm and generative modeling to advance the development of general-purpose intelligent systems. IGM is motivated by the observation that while reinforcement learning (RL) has mastered a wide range of complex simulated tasks, it struggles to generalize in high-dimensional, open-ended tasks. In contrast, generative models excel in such settings due to their expressivity and their ability to serve as powerful priors (e.g., LLMs pretrained on massive corpora). By bridging these two paradigms, IGM offers a promising path forward.

The first direction explored in this dissertation is IGM for …

Irhunter: Universal Detection Of Instruction Reordering Vulnerabilities For Enhanced Concurrency In Distributed And Parallel Systems, Guohua Xin, Guangquan Xu, Yao Zhang, Cheng Wen, Cen Zhang, Xiaofei Xie, Neal N. Xiong, Shaoying Liu, Pan Gao

Research Collection School Of Computing and Information Systems

Instruction reordering is an essential optimization technique used in both compilers and multi-core processors to enhance parallelism and resource utilization. Although the original intent of this technique is to benefit the program, some improper reordering can significantly impact the program correctness, which we call instruction reordering vulnerability (IRV). However, existing methods detect IRV by defining CPU instruction reordering rules to schedule execution paths while neglecting compiler reordering, and thus generate false positives that require manual filtering and resulting in inefficiency. To bridge this gap, in this paper, we propose the IRV detection method, , which analyzes IRV characteristics and extracts …

Contested: Consistency-Aided Tested Code Generation With Llm, Jinhao Dong, Jun Sun, Wenjie Zhang, Jinsong Dong, Dan Hao

Research Collection School Of Computing and Information Systems

Recent advancements in large language models (LLMs) have significantly improved code generation, which generates code snippets automatically based on natural language requirements. Despite achieving state-of-the-art performance, LLMs often struggle to generate accurate and reliable code, requiring developers to spend substantial effort debugging and evaluating the generated output. Researchers have proposed leveraging Consistency to select code that passes more tests (inter-consistency) and demonstrates consistent behavior across more counterparts (intra-consistency). However, since the tests themselves are also generated by LLMs, relying on majority voting based on incorrect tests leads to unreliable results. To address this, we propose a lightweight interaction framework that …

Enhancing Vulnerability Detection Via Inter-Procedural Semantic Completion, Bozhi Wu, Chengjie Liu, Zhiming Li, Yushi Cao, Jun Sun, Shang-Wei Lin

Research Collection School Of Computing and Information Systems

Inspired by advances in deep learning, numerous learning-based approaches for vulnerability detection have emerged, primarily operating at the function level for scalability. However, this design choice has a critical limitation: many vulnerabilities span multiple functions, causing function-level approaches to lose the semantics of called functions and fail to capture true vulnerability patterns. To address this issue, we propose VulnSC, a novel framework designed to enhance learning-based approaches by complementing inter-procedural semantics. VulnSC retrieves the source code of called functions for datasets and leverages large language models (LLMs) with well-designed prompts to generate summaries for these functions. The datasets, enhanced with …

Cashift: Benchmarking Log-Based Cloud Attack Detection Under Normality Shift, Jiongchi Yu, Xiaofei Xie, Qiang Hu, Bowen Zhang, Ziming Zhao, Yun Lin, Lei Ma, Ruitao Feng, Frank Liau

Research Collection School Of Computing and Information Systems

With the rapid advancement of cloud-native computing, securing cloud environments has become an important task. Log-based Anomaly Detection (LAD) is the most representative technique used in different systems for attack detection and safety guarantee, where multiple LAD methods and relevant datasets have been proposed. However, even though some of these datasets are specifically prepared for cloud systems, they only cover limited cloud behaviors and lack information from a whole-system perspective. Another critical issue to consider is normality shift, which implies that the test distribution could differ from the training distribution and highly affect the performance of LAD. Unfortunately, existing works …

Regtrieve: Reducing System-Level Regression Errors For Machine Learning Systems Via Retrieval-Enhanced Ensemble, Junming Cao, Xuwen Xiang, Mingfei Cheng, Bihuan Chen, Xinyan Wang, You Lu, Chaofeng Sha, Xiaofei Xie, Xin Peng

Research Collection School Of Computing and Information Systems

Multiple machine learning (ML) models are often incorporated into real-world ML systems. However, updating an individual model in these ML systems frequently results in regression errors, where the new model performs worse than the old model for some inputs. While model-level regression errors have been widely studied, little is known about how regression errors propagate at system level. To address this gap, we propose RegTrieve, a novel retrieval-enhanced ensemble approach to reduce regression errors at both model and system level. Our evaluation across various model update scenarios shows that RegTrieve reduces system-level regression errors with almost no impact on system …

Unsupervised Recognition Of Unknown Objects For Open-World Object Detection, Ruohuan Fang, Guansong Pang, Wenjun Miao, Xiao Bai, Jin Zheng, Xin Ning

Research Collection School Of Computing and Information Systems

Open-world object detection (OWOD) extends object detection problem to a realistic and dynamic scenario, where a detection model is required to be capable of detecting both known and unknown objects and incrementally learning newly introduced knowledge. Current OWOD models detect the unknowns that exhibit similar features to the known objects, but they suffer from a severe label bias problem, i.e., they tend to detect all regions (including unknown object regions) that are dissimilar to the known objects as part of the background. To eliminate the label bias, this article proposes a novel module, namely reconstruction error-based Weibull (REW) model, that …

Hvi: A New Color Space For Low-Light Image Enhancement, Qingsen Yan, Yixu Feng, Cheng Zhang, Guansong Pang, Kangbiao Shi, Peng Wu, Wei Dong, Jinqiu Sun, Yanning Zhang

Research Collection School Of Computing and Information Systems

Low-Light Image Enhancement (LLIE) is a crucial computer vision task that aims to restore detailed visual information from corrupted low-light images. Many existing LLIE methods are based on standard RGB (sRGB) space, which often produce color bias and brightness artifacts due to inherent high color sensitivity in sRGB. While converting the images using Hue, Saturation and Value (HSV) color space helps resolve the brightness issue, it introduces significant red and black noise artifacts. To address this issue, we propose a new color space for LLIE, namely Horizontal/Vertical-Intensity (HVI), defined by polarized HS maps and learnable inten sity. The former enforces …

Why Does My Transaction Fail? A First Look At Failed Transactions On The Solana Blockchain, Xiaoye Zheng, Zhiyuan Wan, David Lo, Difan Xie, Xiaohu Yang

Research Collection School Of Computing and Information Systems

Solana is an emerging blockchain platform, recognized for its high throughput and low transaction costs, positioning it as a preferred infrastructure for Decentralized Finance (DeFi), Non-Fungible Tokens (NFTs), and other Web 3.0 applications. In the Solana ecosystem, transaction initiators submit various instructions to interact with a diverse range of Solana smart contracts, among which are decentralized exchanges (DEXs) that utilize automated market makers (AMMs), allowing users to trade cryptocurrencies directly on the blockchain without the need for intermediaries. Despite the high throughput and low transaction costs of Solana, the advantages have exposed Solana to bot spamming for financial exploitation, resulting …

Less Is More: On The Importance Of Data Quality For Unit Test Generation, Junwei Zhang, Xing Hu, Shan Gao, Xin Xia, David Lo, Shanping Li

Research Collection School Of Computing and Information Systems

Unit testing is crucial for software development and maintenance. Effective unit testing ensures and improves software quality, but writing unit tests is time-consuming and labor-intensive. Recent studies have proposed deep learning (DL) techniques or large language models (LLMs) to automate unit test generation. These models are usually trained or fine-tuned on large-scale datasets. Despite growing awareness of the importance of data quality, there has been limited research on the quality of datasets used for test generation. To bridge this gap, we systematically examine the impact of noise on the performance of learning-based test generation models. We first apply the open …

Moditector: Module-Directed Testing For Autonomous Driving Systems, Renzhi Wang, Mingfei Cheng, Xiaofei Xie, Yuan Zhou, Lei Ma

Research Collection School Of Computing and Information Systems

Testing Autonomous Driving Systems (ADSs) is crucial for ensuring their safety, reliability, and performance. Despite numerous testing methods available that can generate diverse and challenging scenarios to uncover potential vulnerabilities, these methods often treat ADS as a black-box, primarily focusing on identifying system-level failures like collisions or near-misses without pinpointing the specific modules responsible for these failures. This lack of root causes understanding for the failures hinders effective debugging and subsequent system repair. Furthermore, current approaches often fall short in generating violations that adequately test the individual modules of an ADS from a system-level perspective, such as perception, prediction, planning, …

Demystifying Memorization In Llm-Based Program Repair Via A General Hypothesis Testing Framework, Jiaolong Kong, Xiaofei Xie, Shangqing Liu

Research Collection School Of Computing and Information Systems

Large Language Models (LLMs) have achieved remarkable success in various applications, particularly in code-related tasks such as code generation and program repair, setting new performance benchmarks. However, the extensive use of large training corpora raises concerns about whether these achievements stem from genuine understanding or mere memorization of training data—a question often overlooked in current research. This paper aims to study the memorization issue within LLM-based program repair by investigating whether the correct patches generated by LLMs are the result of memorization. The key challenge lies in the absence of ground truth for confirming memorization, leading to various ad-hoc methods …