Digital Commons Network^™

Trash Bot!, Owen Hoffman , ‘26, Ryder Maston , ‘26, Jack Nicholson , ‘26

Senior Theses, Projects, and Awards

Trash Bot! is the prototype of a smart and mobile waste receptacle for Swarthmore’s waste management system. With Swarthmore College committing to embed zero waste into campus culture and achieve a 90% diversion rate by 2035, flexible waste systems must be developed and improved. Trash Bot! seeks to improve the diversion rate and sort accuracy on campus by reducing the human error present in the current waste sorting process. Trash Bot! was also designed to seamlessly integrate with a classroom environment, maintaining safety standards and avoiding collisions. Trash Bot! uses computer vision coupled with a trained classifier algorithm to achieve …

Generative Ai In Enterprises: Optimizing Applications With Large Language Models, Donghao Huang

Dissertations and Theses Collection (Open Access)

This dissertation investigates how to deploy Large Language Models (LLMs) effectively in enterprise settings, where accuracy, reliability, cost, privacy, and operational constraints often matter more than benchmark performance alone. Drawing on seventeen peer-reviewed publications (eleven published and six accepted for publication), the work develops and validates optimization strategies across three connected themes: retrieval-augmented generation (RAG), agentic AI for workflow automation, and deployment guidelines for real-world enterprise environments.

First, we study RAG optimization through systematic evaluation of open and proprietary models, highlighting conditions under which efficient open-weight models can match or exceed proprietary alternatives. To address a pervasive failure mode in …

Explainable Artificial Intelligence In The Image Domain And Its Applications To The Medical Field, Mirtha Lucas

Theses and Dissertations from DePaul University

This dissertation investigates the development of Explainable Artificial Intelligence (XAI) methods for deep learning models in the image domain, with a particular focus on medical imaging applications. Although neural networks achieve high predictive performance, their lack of interpretability limits their adoption in critical domains such as healthcare, where transparency and trust are essential. This work addresses this challenge by proposing novel approaches that improve the interpretability and reliability of model predictions.   A primary contribution is the introduction of Riemann–Stieltjes Integrated Grad-CAM (RSI Grad-CAM), a gradient-based attribution method that generates more relevant and spatially localized saliency maps. The method is evaluated …

Agentspec: Customizable Runtime Enforcement For Safe And Reliable Llm Agents, Haoyu Wang, Christopher M. Poskitt, Jun Sun

Research Collection School Of Computing and Information Systems

Agents built on LLMs are increasingly deployed across diverse domains, automating complex decision-making and task execution. However, their autonomy introduces safety risks, including security vulnerabilities, legal violations, and unintended harmful actions. Existing mitigation methods, such as model-based safeguards and early enforcement strategies, fall short in robustness, interpretability, and adaptability. To address these challenges, we propose AgentSpec, a lightweight domain-specific language for specifying and enforcing runtime constraints on LLM agents. With AgentSpec, users define structured rules that incorporate triggers, predicates, and enforcement mechanisms, ensuring agents operate within predefined safety boundaries. We implement AgentSpec across multiple domains, including code execution, embodied agents, …

Hypersiniel: Guaranteed Output Delivery Comes (Almost) Free In Private Delegation Of Zksnarks, Yunbo Yang, Yuejia Cheng, Junkai Liang, Kailun Wang, Xuanming Liu, Xiaoguo Li, Jianfei Sun, Jiachen Shen, Xiaolei Dong, Zhenfu Cao, Meng Hao, Guomin Yang, Deng, Robert H., Kui Ren

Research Collection School Of Computing and Information Systems

Zero-knowledge Succinct Non-interactive Argument of Knowledge (zkSNARK) is a powerful cryptographic primitive that enables a prover to convince a verifier that something is true without leaking the private witness.Current zkSNARKs face significant computational costs in generating proofs, which restricts their use in areas like private payments, confidential smart contracts, and anonymous credentials. Private delegation offers a practical solution by outsourcing the heavy computation to powerful external workers without leaking any private information. In this work, we propose HyperSiniel, an efficient private delegation framework for general zkSNARKs that achieves a new feature called guaranteed output delivery (GOD). HyperSiniel is designed to …

Bridging Draft Policy Misalignment: Group Tree Optimization For Speculative Decoding, Shijing Hu, Jingyang Li, Zhihui Lu, Pan Zhou

Research Collection School Of Computing and Information Systems

Speculative decoding accelerates large language model (LLM) inference by letting a lightweight draft model propose multiple tokens that the target model verifies in parallel. Yet existing training objectives optimize only a single greedy draft path, while decoding follows a tree policy that re-ranks and verifies multiple branches. This draft policy misalignment limits achievable speedups. We introduce Group Tree Optimization (GTO), which aligns training with the decoding-time tree policy through two components: (i) Draft Tree Reward, a sampling-free objective equal to the expected acceptance length of the draft tree under the target model, directly measuring decoding performance; (ii) Group-based Draft Policy …

From Spatial To Actions: Grounding Vision-Language-Action Model In Spatial Foundation Priors, Zhengshen Zhang, Hao Li, Yalun Dai, Zhengbang Zhu, Lei Zhou, Chenchen Liu, Dong Wang, Francis E. H. Tay, Sijin Chen, Ziwei Liu, Yuxiao Liu, Xinghang Li, Pan Zhou

Research Collection School Of Computing and Information Systems

Existing vision-language-action (VLA) models act in 3D real-world but are typically built on 2D encoders, leaving a spatial reasoning gap that limits generalization and adaptability. Recent 3D integration techniques for VLAs either require specialized sensors and transfer poorly across modalities, or inject weak cues that lack geometry and degrade vision-language alignment. In this work, we introduce FALCON (From Spatial to Action), a novel paradigm that injects rich 3D spatial tokens into the action head. FALCON leverages spatial foundation models to deliver strong geometric priors from RGB alone, and includes an Embodied Spatial Model that can optionally fuse depth, or pose …

Ai Interpretability In Healthcare Communication, Ananya Jeyappragash

Dartmouth College Master’s Theses

Artificial intelligence has increasingly been adopted in healthcare, largely for specialized tasks and under significant human oversight. The use of large black-box systems raises important concerns about transparency in high-stakes environments such as clinical decision-making. Clinical communication is fundamentally human-centered, and failures in judgment can have serious consequences for patient care. Overestimating the reasoning abilities of large language models may lead to undue trust in fabricated or “hallucinated” outputs, while rejecting AI-assisted tools altogether may preserve inefficient workflows and contribute to missed or delayed diagnoses. These concerns reflect a broader tradeoff between accuracy and interpretability: although more complex models may …

Aiw26s: Machine Learning Of Structured Data, Moumita Saha

Paul English Applied Artificial Intelligence (AI) Institute Publications

This workshop introduces the fundamentals of machine learning for structured data, focusing on tabular datasets and real-world applications. Participants explore key concepts such as data types, data preprocessing, feature engineering, and supervised learning methods. The session covers commonly used models, including linear regression, logistic regression, decision trees, and neural networks, along with evaluation metrics such as RMSE, accuracy, and confusion matrices. By the end of the workshop, participants will have gained a practical understanding of how to build, interpret, and evaluate machine learning models for structured data.

Knowledge Distillation From A Large Vision-Language Model To Compact Students For Architectural Floor Plan Understanding, Kiran Silwal

Honors Theses

In this research, the use of a large vision-language model to train smaller, deployable models for architectural floor plan question answering is investigated. Reading a floor plan today requires either a human expert or a paid query to a proprietary model, and neither option is practical for real-estate platforms that must process thousands of units at scale. To address this problem, a knowledge distillation approach is employed in which a large teacher model (GPT-4.1-mini) generates labeled question-answer pairs from floor plan images, and smaller student models learn from those labels. The teacher produced 37,027 labeled pairs from 12,343 floor plan …

Ai-Assisted 3d Pre-Visualization: Streamlining Animatics For Filmmakers, Nishit Thapa

Honors Theses

Pre-visualization is a core part of film pre-production. Creators use it to test blocking, lighting, camera movement, and spatial relationships before committing to a shot. Despite this, three-dimensional pre-visualization remains out of reach for many entry-level filmmakers, as the software is expensive and takes substantial time to learn.

Recent AI developments have produced generative systems capable of creating highly re-alistic imagery and video from text prompts. These systems open new possibilities for visual sto-rytelling but are not built for the iterative, hands-on process that pre-visualization requires, where creators must adjust scene parameters, camera logic, and spatial configurations throughout.

This thesis …

Rethinking News Classification Through A Multi-Dimensional Framework, Luana De Jesus Ferreira

Honors Theses

This thesis proposes a multi-dimensional framework for news classification that evaluates articles across three independent dimensions: headline accuracy, language neutrality, and content reliability. These dimensions produce both a continuous reliability score and a five-tier interpretive scale, while additionally classifying articles by genre and topic. To operationalize this framework, a structured annotation protocol was developed and applied to a dataset of 373 news articles drawn from 79 outlets spanning a wide range of contemporary media ecosystem. A binary Logistic Regression classifier trained on the ISOT Fake News Dataset was then evaluated against this dataset to examine how a model trained on …

Generative Artificial Intelligence With A Human Touch: Building Hana, Conrad Johnson

Faculty Scholarship

This Essay examines how generative artificial intelligence (GenAI) can be integrated into legal education and public interest law practice in a way that meaningfully enhances — rather than diminishes — human judgment, professional responsibility, and access to justice. Drawing on the experience of Columbia Law School’s Lawyering in the Digital Age Clinic, the Essay situates GenAI within an experiential pedagogy that emphasizes competence, ethical awareness, and collaborative problem-solving. It argues that law students and lawyers must move beyond a passive or uncritical use of GenAI tools; toward a deeper understanding of how these systems operate, the risks they pose, and …

Exploring The U.S. Public Service Employees’ Experiences With Gamified Cybersecurity Awareness Training, Bisola Adepoju

Human Resource Development Theses and Dissertations

The purpose of this qualitative study was to explore the U.S. public service employees' experiences with gamified cybersecurity awareness training. The primary research question guiding this inquiry was: How do U.S. public service employees make sense of their lived experiences with gamified cybersecurity awareness training? Two secondary questions examined what aspects of the training employees perceive as meaningful or disengaging, and how they interpret their motivation, engagement, and cybersecurity awareness during training. I chose an interpretive qualitative research design grounded in a constructivist paradigm and purposefully selected 11 U.S. public service employees who had participated in gamified cybersecurity awareness training. …

Federated Retrieval-Augmented Generation For Cybersecurity In Resource-Constrained Iot And Edge Environments: A Deployment-Oriented Scoping Review, Hangyu He, Yuan, Kai Wu, Wei Ni

Research outputs 2022 to 2026

Cybersecurity operations in IoT and edge environments require fast, evidence-grounded decisions under strict resource and trust constraints. While large language models can support triage and incident analysis, their parametric knowledge may be outdated and prone to hallucination. Retrieval-augmented generation (RAG) improves grounding by conditioning responses on retrieved evidence, but also introduces new risks such as knowledge-base poisoning, indirect prompt injection, and embedding leakage. Federated learning enables collaborative adaptation without centralizing sensitive data, motivating federated RAG (FedRAG) architectures for distributed cybersecurity deployments. This study presents a deployment-oriented scoping review of FedRAG for cybersecurity. The review follows PRISMA-ScR reporting guidance and synthesizes …

Conflogger: Enhance Systems’ Configuration Diagnosability Through Configuration Logging, Shiwen Shan, Yintong Huo, Yuxin Su, Zhining Wang, Dan Li, Zibin Zheng

Research Collection School Of Computing and Information Systems

Modern configurable systems offer customization via intricate configuration spaces, yet such flexibility introduces pervasive configuration-related issues such as misconfigurations and latent softwarebugs. Existing diagnosability supports focus on post-failure analysis of software behavior to identify configuration issues, but none of these approaches look into whether the software clue sufficient failure information for diagnosis. To fill in the blank, we propose the idea of configuration logging to enhance existing logging practices at the source code level. We develop ConfLogger, the first tool that unifies configuration-aware static taint analysis with LLM-based log generation to enhance software configuration diagnosability. Specifically, our method 1) identifies …

Propaganda Ai: An Analysis Of Semantic Divergence In Large Language Models, Nay Myat Min, Long H. Pham, Yige Li, Jun Sun

Research Collection School Of Computing and Information Systems

Large language models (LLMs) can exhibit concept-conditioned semantic divergence: common high-level cues (e.g., ideologies, public figures) elicit unusually uniform, stance-like responses that evade token-trigger audits. This behavior falls in a blind spot of current safety evaluations, yet carries major societal stakes, as such concept cues can steer content exposure at scale. We formalize this phenomenon and present RAVEN (Response Anomaly Vigilance), a black-box audit that flags cases where a model is simultaneously highly certain and atypical among peers by coupling semantic entropy over paraphrastic samples with cross-model disagreement. In a controlled LoRA fine-tuning study, we implant a concept-conditioned stance using …

Where Did It Go Wrong? Attributing Undesirable Llm Behaviors Via Representation Gradient Tracing, Zhe Li, Wei Zhao, Yige Li, Jun Sun

Research Collection School Of Computing and Information Systems

Large Language Models (LLMs) have demonstrated remarkable capabilities, yet their deployment is frequently undermined by undesirable behaviors such as generating harmful content, factual inaccuracies, and societal biases. Diagnosing the root causes of these failures poses a critical challenge for AI safety. Existing attribution methods, particularly those based on parameter gradients, often fall short due to prohibitive noisy signals and computational complexity. In this work, we introduce a novel and efficient framework that diagnoses a range of undesirable LLM behaviors by analyzing representation and its gradients, which operates directly in the model's activation space to provide a semantically meaningful signal linking …

Reducing Class-Wise Performance Disparity Via Margin Regularization, Beier Zhu, Kesen Zhao, Jiequan Cui, Qianru Sun, Yuan Zhou, Xun Yang, Hanwang Zhang

Research Collection School Of Computing and Information Systems

Deep neural networks often exhibit substantial disparities in class-wise accuracy, even when trained on class-balanced data—posing concerns for reliable deployment. While prior efforts have explored empirical remedies, a theoretical understanding of such performance disparities in classification remains limited. In this work, we present Margin Regularization for performance disparity Reduction (MR2 ), a theoretically principled regularization for classification by dynamically adjusting margins in both the logit and representation spaces. Our analysis establishes a margin-based, class-sensitive generalization bound that reveals how per-class feature variability contributes to error, motivating the use of larger margins for “hard” classes. Guided by this insight, MR2 optimizes …

Llmqua: Practical Backdoor Injection On Large Language Model Quantization, Xiangxiang Chen, Peixin Zhang, Jun Sun, Jin Song Dong, Wenhai Wang, Jingyi Wang

Research Collection School Of Computing and Information Systems

Quantization is widely used to enable local deployment of large language models (LLMs) on resource-constrained devices. Recent work (e.g., QuRA) shows quantization can be exploited via rounding manipulation to implant backdoors. However, such an attack has been evaluated only on small models and does not directly apply to LLMs due to three key constraints: (1) limited poisoning data from small, task-agnostic calibration sets; (2) layer-wise quantization restricting adversarial access to global representations; and (3) lack of gradient access in quantization pipelines, blocking gradient-based attacks.We propose LLMQuA, a practical quantization-phase backdoor attack tailored to the LLM setting. LLMQuA (i) injects backdoors …

Prompting Frameworks For Large Language Models: A Survey, Xiaoxia Liu, Jingyi Wang, Jun Sun, Xiaohan Yuan, Guoliang Dong, Peng Di, Wenhai Wang, Dongxia Wang

Research Collection School Of Computing and Information Systems

Since the launch of ChatGPT, a powerful AI Chatbot developed by OpenAI, large language models (LLMs) have made significant advancements in both academia and industry, bringing about a fundamental engineering paradigm shift in many areas. While LLMs are powerful, it is also crucial to best use their power where “prompt” plays a core role. However, the booming LLMs themselves, including excellent APIs like ChatGPT, have several inherent limitations: (1) temporal lag of training data, and (2) the lack of physical capabilities to perform external actions. Recently, we have observed the trend of utilizing prompt-based tools to better utilize the power …

Who You Explain To Matters: Learning By Explaining To Conversational Agents With Different Pedagogical Roles, Zhengtao Xu, Junti Zhang, Anthony Tang, Yi-Chieh Lee

Research Collection School Of Computing and Information Systems

Conversational agents are increasingly used in education for learning support. An application is “learning by explaining”, where learners explain their understanding to an agent. However, existing research focuses on single roles, leaving it unclear how different pedagogical roles influence learners’ interaction patterns, learning outcomes and experiences. We conducted a between-subjects study (N=96) comparing agents with three pedagogical roles (Tutee, Peer, Challenger) and a control condition while learning an economics concept. We found that different pedagogical roles shaped learning dynamics, including interaction patterns and experiences. Specifically, the Tutee agent elicited the most cognitive investment but led to high pressure. The Peer …

Learning Feature Inversion For Multi-Class Anomaly Detection Under General-Purpose Coco-Ad Benchmark, Jiangning Zhang, Chengjie Wang, Xiangtai Li, Guanzhong Tian, Zhucun Xue, Yong Liu, Guansong Pang, Dacheng Tao

Research Collection School Of Computing and Information Systems

Anomaly detection (AD) is often focused on detecting anomaly areas for industrial quality inspection and medical lesion examination. However, due to the specific scenario targets, the data scale for AD is relatively small, and evaluation metrics are still deficient compared to classic vision tasks, such as object detection and semantic segmentation. To fill these gaps, this work first constructs a large-scale and general-purpose COCO-AD dataset by extending COCO to the AD field. This enables fair evaluation and sustainable development for different methods on this challenging benchmark. Moreover, current metrics such as AU-ROC have nearly reached saturation on simple datasets, which …

Semat: Semantic Enhanced Natural Image Interactive Matting, Ruihao Xia, Yu Liang, Peng-Tao Jiang, Hao Zhang, Qianru Sun, Yang Tang, Bo Li, Pan Zhou

Research Collection School Of Computing and Information Systems

Recent approaches attempt to adapt powerful interactive segmentation models, such as SAM, to interactive matting and fine-tune the models based on synthetic matting datasets. However, models trained on synthetic data fail to generalize to complex and occlusion scenes. We address this challenge by proposing a new matting dataset based on the COCO dataset, namely COCO-Matting. It selects real-world complex images from COCO and converts semantic segmentation masks to matting labels. The built COCO-Matting comprises an extensive collection of 36,980 human instance-level alpha mattes in complex natural scenarios. Furthermore, existing SAM-based matting methods extract intermediate features and masks from a frozen …

Visual Loop: Bridging The Cognitive Gap In Software Development Through Visual-Ai Collaboration, Luis Filipe Fernandes Gomes, Xin Zhou, David Lo, Rui Abreu

Research Collection School Of Computing and Information Systems

Software development remains predominantly text-centric, despite decades of evidence showing that developers think and communicate visually. While sketches and diagrams externalize developers’ mental models, they remain disconnected from source code and quickly become outdated. Recent advances in foundation models, capable of both code and visual reasoning, create an opportunity to unify these representations. In this vision paper, we introduce Visual Loop, a continuous visual development environment that keeps code and informal sketches in bidirectional synchronization. Our prototype connects a code editor with a tablet-based visualization workspace, allowing developers to explore, annotate, and modify systems through freehand sketches interpreted by multimodal …

On Autopilot? An Empirical Study Of Human-Ai Teaming And Review Practices In Open Source, Haoyu Gao, Peerachai Banyongrakkul, Hao Guan, Mansooreh Zahedi, Christoph Treude

Research Collection School Of Computing and Information Systems

Large Language Models (LLMs) increasingly automate software engineering tasks. While recent studies highlight the accelerated adoption of “AI as a teammate” in Open Source Software (OSS), developer interaction patterns remain under-explored. In this work, we investigated project-level guidelines and developers’ interactions with AI-assisted pull requests (PRs) by expanding the AIDev dataset to include finer-grained contributor code ownership and a comparative baseline of human-created PRs. We found that over 67.5% of AI-co-authored PRs originate from contributors without prior code ownership. Despite this, the majority of repositories lack guidelines for AI-coding agent usage. Notably, we observed a distinct interaction pattern: AI-co-authored PRs …

Structural Silence: When Ai Infrastructure Fails Speakers Of Underrepresented Languages, Avijit Roy, Proma Roy

Publications and Research

Artificial intelligence tools for education and language support are increasingly framed as scalable responses to access gaps in under-resourced communities. Yet the infrastructure underlying these tools—training corpora, tokenization schemes, evaluation benchmarks, and deployment architectures—encodes a set of assumptions that systematically disadvantages speakers of underrepresented languages before a single model is trained. This paper examines those assumptions through the lens of Bengali, one of the world’s most widely spoken languages with roughly 285 million speakers (Ethnologue, 2025; International Communication and Leadership School, 2026), and the structural barriers that emerge when attempting to build AI-assisted educational tools for Bengali-speaking learners in low-connectivity …

Portrait Shadow Removal Via Self-Exemplar Illumination Equalization, Qian Huang, Cheng Xu, Guiqing Li, Ziheng Wu, Shengxin Liu, Shengfeng He

Research Collection School Of Computing and Information Systems

We introduce the Self-Exemplar Illumination Equalization Network, designed specifically for effective portrait shadow removal. The core idea of our method is that partially shadowed portraits can find ideal exemplars within their non-shadowed facial regions. Rather than directly fusing two distinct classes of facial features, our approach utilizes non-shadowed regions as an illumination indicator to equalize the shadowed regions, generating deshadowed results without boundary-merging artifacts. Our network comprises cascaded Self-Exemplar Illumination Equalization Blocks (SExmBlock), each containing two modules: a self-exemplar feature matching module and a feature-level illumination rectification module. The former identifies and applies internal illumination exemplars to shadowed areas, producing …

Patchgpt: Multi-Agent Patch Backporting Without Model Fine-Tuning, Ye Liu, Ruidong Han, Chengyan Ma, Yuqing Niu, David Lo

Research Collection School Of Computing and Information Systems

Patch backporting is crucial and prevalent in the maintenance of modern open-source software such as Linux kernels and forked repositories. However, porting patches across program versions remains a challenging problem due to the complexity of synergizing diverse patches with divergent program versions. In this paper, we propose PatchGPT, an agentic patch backporting framework for fine-grained patch generation. PatchGPT encompasses three agents: Miner for decomposing a sequence of atomic change steps as the original patch plan, Adapter for adapting the patch plan, and Executor for executing the adapted patch plan according to predefined change semantics. We conduct experiments on the PPatHF’s …

Understanding Codebase Like A Professional! Human-Ai Collaboration For Code Comprehension, Jie Gao, Yue Xue, Xiaofei Xie, Junming Cao, Soemin Thant, Erika Lee, Bowen Xu

Research Collection School Of Computing and Information Systems

Understanding an unfamiliar codebase is an essential task for developers in various scenarios, such as during the onboarding process. Especially when the codebase is large and time is limited, achieving a decent level of comprehension remains challenging for both experienced and novice developers, even with the assistance of large language models (LLMs). Existing studies have shown that LLMs often fail to support users in understanding code structures or to provide user-centered, adaptive, and dynamic assistance in real-world settings.To address this, we propose learning from the perspective of a unique role, code auditors, whose work often requires them to quickly familiarize …