Physical Sciences and Mathematics Commons^™

Stacked From One: Multi-Scale Self-Injection For Context Window Extension, Wei Han, Pan Zhou, Shuicheng Yan

Research Collection School Of Computing and Information Systems

The limited context window of contemporary large language models (LLMs) remains a primary bottleneck for their broader application across diverse domains. Although continual pre-training on long-context data offers a straightforward solution, it incurs prohibitive data acquisition and computational costs. To address this challenge, we propose SHAREDLLM, a novel framework based on multi-grained context compression and query-aware information acquisition. SHAREDLLM comprises two stacked short-context LLMs: a lower model serving as a compressor and an upper model acting as a decoder. The lower model compresses long inputs into compact, multi-grained representations, which are then forwarded to the upper model for context-aware processing. …

Dragging With Geometry: From Pixels To Geometry-Guided Image Editing, Xinyu Pu, Hongsong Wang, Jie Gui, Pan Zhou

Research Collection School Of Computing and Information Systems

Interactive point-based image editing serves as a controllable editor, enabling precise and flexible manipulation of image content. However, most drag-based methods operate primarily on the 2D pixel plane with limited use of 3D cues. As a result, they often produce imprecise and inconsistent edits, particularly in geometry-intensive scenarios such as rotations and perspective transformations. To address these limitations, we propose a novel geometry-guided drag-based image editing method—GeoDrag, which addresses three key challenges: 1) incorporating 3D geometric cues into pixel-level editing, 2) mitigating discontinuities caused by geometry-only guidance, and 3) resolving conflicts arising from multi-point dragging. Built upon a unified displacement …

Dreamcs: Geometry-Aware Text-To-3d Generation With Unpaired 3d Reward Supervision, Xiandong Zou, Ruihao Xia, Hongsong Wang, Pan Zhou

Research Collection School Of Computing and Information Systems

While text-to-3D generation has attracted growing interest, existing methods often struggle to produce 3D assets that align well with human preferences. Current preference alignment techniques for 3D content typically rely on hardly-collected preference-paired multi-view 2D images to train 2D reward models, when then guide 3D generation — leading to geometric artifacts, such as the Janus face problem and geometric incompleteness, due to their inherent 2D bias. To address these limitations, we construct 3D-MeshPref, the first large-scale unpaired 3D preference dataset, featuring diverse 3D meshes annotated by a large language model and refined by human evaluators. We then develop RewardCS, the …

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 …

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 …

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 …

Managing Reproducibility Debt In Scientific Software: A Practical Framework, Zara Hassan, Christoph Treude, Graham Williams, Michael Norrish, Alex Potanin

Research Collection School Of Computing and Information Systems

Scientific software includes end-user applications, modelling tools, research software for publications, and production systems for real users. It plays a key role across various scientific disciplines by enabling large-scale computation, simulation, and data analysis. Unlike commercial software, scientific software is often developed in dynamic research environments with limited engineering practices, documentation, or testing. This makes it fragile and difficult to reproduce results, even when code and data are available, conditions in which Reproducibility Debt (RpD) accumulates. This paper presents the Reproducibility Debt Management Framework (RpD-MF), which is grounded in evidence from a systematic literature review, practitioner interviews, and a global …

Real-Time Motion-Controllable Autoregressive Video Diffusion, Kesen Zhao, Jiaxin Shi, Beier Zhu, Junbao Zhou, Xiaolong Shen, Yuan Zhou, Qianru Sun, Hanwang Zhang

Research Collection School Of Computing and Information Systems

Real-time motion-controllable video generation remains challenging due to the inherent latency of bidirectional diffusion models and the lack of effective autoregressive (AR) approaches. Existing AR video diffusion models are limited to simple control signals or text-to-video generation, and often suffer from quality degradation and motion artifacts in few-step generation. To address these challenges, we propose AR-Drag, the first RL-enhanced few-step AR video diffusion model for real-time image-to-video generation with diverse motion control. We first fine-tune a base I2V model to support basic motion control, then further improve it via reinforcement learning with a trajectory-based reward model. Our design preserves the …

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 …

Be Responsible In Your Answers! Monitoring Out-Of-Domain Behaviors In Domain-Specific Llms, Boquan Li, Chenzhe Lou, Zhe Ren, Peixin Zhang, Zirui Fu, Jun Sun, Yaowen Zheng

Research Collection School Of Computing and Information Systems

Large Language Models (LLMs) have accelerated the rapid development of chatbot web applications in various domains, such as coding, biomedicine and psychology. Compared to general LLMs like ChatGPT, domain-specific LLMs require a greater sense of responsibility. For instance, if a programming LLM casually answers medical or psychological questions, it not only misleads the public but also poses legal risks. This highlights new demands for monitoring and preventing such irresponsible behaviors. Existing efforts attempt to monitor LLMs from multiple aspects, such as lying, jailbreaks, and toxic content, while overlooking out-of-domain behaviors. In this work, we propose an innovative LLM domain monitoring …

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 …

Teamwise: Exploring Virtually Embodied Ai Facilitation For Video-Based Team Onboarding, Venkata Akhila Rani Obilisetty, Mikkeline Elleby, Anthony Tang, April Yi Wang

Research Collection School Of Computing and Information Systems

AI-mediated facilitation has emerged as a scalable approach to supporting onboarding and coordination in newly formed remote teams, yet existing systems are predominantly text-based. To explore how video-based, virtually embodied AI facilitators shape team experiences, we present TeamWise, which joins video-based onboarding meetings as an on-screen avatar. TeamWise guides teams through a structured facilitation flow of low-stakes activities to foster rapport, mutual awareness, and shared identity. While the overall sequence of activities and facilitation goals is predefined, the facilitator’s turn-by-turn utterances are generated dynamically by an LLM in response to participant input. We conducted a formative study of TeamWise to …

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 …

Challenges In Synchronous And Remote Collaboration Around Visualization, Matthew Brehmer, Maxime Cordeil, Christophe Hurter, Takayuki Itoh, Wolfgang Büschel, Mahmood Jasim, Arnaud Prouzeau, David Saffo, Lyn Bartram, Sheelagh Carpendale, Chen Zhu-Tian, Andrew Cunningham, Anthony Tang, Samuel Huron, Masahiko Itoh, Arpit Joshi, Kiyoshi Kiyokawa, Hideaki Kuzuoka, Bongshin Lee, Guillermo Molina León

Research Collection School Of Computing and Information Systems

We characterize 16 challenges faced by those investigating and developing remote and synchronous collaborative experiences around visualization. Our work reflects the perspectives and prior research efforts of an international group of 29 experts from across human-computer interaction and visualization sub-communities. The challenges are anchored around five collaborative activities that exhibit a centrality of visualization and multimodal communication. These activities include exploratory data analysis, creative ideation, visualization-rich presentations, joint decision making grounded in data, and real-time data monitoring. The challenges also reflect the changing dynamics of these activities in the face of recent advances in extended reality (XR) and artificial intelligence …

Reasoning On Time-Series For Financial Technical Analysis, Kelvin J. L. Koa, Jan Chen, Yunshan Ma, Huanhuan Zheng, Tat-Seng Chua

Research Collection School Of Computing and Information Systems

While Large Language Models have been used to produce interpretable stock forecasts, they mainly focus on analyzing textual reports but not historical price data, also known as Technical Analysis. This task is challenging as it switches between domains: the stock price inputs and outputs lie in the time-series domain, while the reasoning step should be in natural language. In this work, we introduce Verbal Technical Analysis (VTA), a novel framework that combine verbal and latent reasoning to produce stock time-series forecasts that are both accurate and interpretable. To reason over time-series, we convert stock price data into textual annotations and …

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 …

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 …

Autobackdoor: Automating Backdoor Attacks Via Llm Agents, Yige Li, Zhe Li, Wei Zhao, Nay Myat Min, Hanxun Huang, Xingjun Ma, Jun Sun

Research Collection School Of Computing and Information Systems

Backdoor attacks pose a serious threat to the secure deployment of large language models (LLMs), enabling adversaries to implant hidden behaviors triggered by specific inputs. However, existing methods often rely on manually crafted triggers and static data pipelines, which are rigid, labor-intensive, and inadequate for systematically evaluating modern defense robustness. As AI agents become increasingly capable, there is a growing need for more rigorous, diverse, and scalable \textit{red-teaming frameworks} that can realistically simulate backdoor threats and assess model resilience under adversarial conditions. In this work, we introduce \textsc{AutoBackdoor}, a general framework for automating backdoor injection, encompassing trigger generation, poisoned data …

Ultra-High Vacuum Outgassing Characterization Of Thermally Processed Low-Carbon Steel For Advanced Particle Accelerator And Gravitational Wave Detector Applications, Aiman H. Al-Allaq

Mechanical & Aerospace Engineering Theses & Dissertations

This dissertation investigated AISI 1020 low-carbon steel as an alternative vacuum chamber material to conventional stainless steel for ultra-high vacuum (UHV) and extreme-high vacuum (XHV) applications. After a 400 °C/48 h bake, AISI 1020 tube chambers achieved a hydrogen outgassing rate of 2.4 × 10⁻¹⁶ Torr·L·s⁻¹·cm⁻², approximately 2,300 times lower than the prebaked 316L stainless-steel comparator, among the lowest hydrogen outgassing rates ever reported for an uncoated metallic vacuum chamber. Bare and magnetite-coated AISI 1020 chambers were then compared using throughput and rate-of-rise methods. The magnetite coating yielded 5× lower water outgassing at room temperature, but this advantage disappeared after …

Dissecting The Etiology Of Alcohol Use Disorder By An Integrative Heritable Component Approach, Ivy Garrenton

Computer Science Theses & Dissertations

Alcohol Use Disorder (AUD) is a pervasive condition characterized by complex interplay among genetic, phenotypic, and environmental factors. Although previous studies have identi fied genetic loci associated with alcohol consumption, these efforts have not captured the genetic heterogeneity and gene-environment interactions underlying AUD pathogenesis. To address this critical gap, we developed a novel statistical methodology that integrates phenotypic, genotypic, and environmental data through an environmentally modified Genetic Relationship Matrix (GRM) to derive AUD-related traits with enhanced heritability.

This approach demonstrated superior performance in both simulated and real-world datasets. Traits derived using the environmentally modified GRM exhibited significantly higher estimated heritability …

Competitive Reaction Pathways In The Fully Green Synthesis And Surface Modification Of Silver Nanowires Under L-Ascorbic Acid-Limited Conditions, Finley Neal

Honors Theses

Silver nanowires (AgNWs) are critical building blocks for transparent conductive films (TCFs) in flexible electronics, yet their commercial viability is hindered by environmentally hazardous, energy-intensive synthesis methods and limited intrinsic chemical stability. This study investigated a fully green, room-temperature (25℃) batch synthesis of pristine AgNWs using tannic acid as a biodegradable dual reducing and capping agent under strict kinetic pH control (pH 1.25). To enhance oxidative resistance, a subsequent surface modification protocol was evaluated to deposit a protective palladium (Pd) shell using L-ascorbic acid (LAA) as a secondary reducing agent. The pristine green synthesis yielded one-dimensional Ag nanostructures with a …

Synthesis And Photophysical Studies Of Coumarin-Quinoline Derivatives, Nishrenika Pokhrel

Honors Theses

A new family of oxime-functionalized coumarin-quinoline derivatives, 5a and 5b, has been synthesized through a five-step procedure in high yields (45-98%). Compound 5a is designed as the sensing probe, while 5b serves as a model system to support mechanistic understanding. The sensing approach is based on the reactivity of the oxime group, which can undergo phosphorylation in the presence of organophosphate nerve agents. Based on this behavior, it is hypothesized that 5a can react irreversibly with a nerve agent mimic. The structures were fully characterized by NMR, FTIR, and HRMS. Photophysical properties of 5a were studied in various solvents, …

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 …

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 …

Discovery Of A New Cataclysmic Variable, Galex J211240.9–422959, Via The Exoasteroids Citizen Science Project, Mei Lin, Aaron Meisner, Siyi Xu, Arttu Sainio, Liliana E. Rivera Sandoval, Paula Szkody, Christopher Tanner, Laura K. Rogers, Sarah Casewell, Austin Humphreys

Physics & Astronomy Faculty Publications

We report the discovery and spectroscopic classification of a new cataclysmic variable (CV), GALEX J211240.9–422959 (Gaia DR3 source ID 6580244356829724928). This object was initially identified as a source exhibiting strong mid-infrared variability in Wide-field Infrared Survey Explorer (WISE) time-series data by a volunteer participating in the Exoasteroids citizen science project. Prompted by this strong variability at 3–5 μm, we obtained a follow-up optical spectrum using the Gemini Multi-Object Spectrograph at the Gemini South Observatory. The resulting spectrum displays a blue continuum dominated by strong, broad emission lines of the hydrogen Balmer series and neutral helium, spectral signatures characteristic of …

Antimicrobial Performance Of Amorphous Vs. Crystalline La1-Xsrxcoo3 Perovskites, S. M. Au Shohag, Fahmida Akhter, Ahmed Touhami, Swati Mohan, Md. Wasikur Rahman, Mohammed Jasim Uddin

Physics & Astronomy Faculty Publications

Perovskite materials, with their unique crystal structure and chemical properties, are emerging as promising antimicrobial agents for potential biomedical applications. In this research, different weight percentages of Sr2 + doped amorphous and crystalline Lanthanum Cobaltite (La1-xSrxCoO3 defined as LSCO) (x = 0, 0.025, 0.05, 0.1, and 0.15) perovskite was synthesized by using a simple combination of sol-gel and molten-salt processes and characterized by X-ray diffraction (XRD), and scanning electron microscopy (SEM) ensures the uniform morphology of the materials. During antimicrobial study using the Kirby Bauer method, amorphous LSCO showed much better activity than crystalline LSCO against Escherichia coli, Vibrio …

A Search For Magnetically Active Binary Counterparts To Chandra X-Ray Sources In 47 Tucanae With Hst Imaging, Haldan N. Cohn, Phyllis M. Lugger, Craig O. Heinke, Maureen Van Den Berg, Liliana Rivera Sandoval, Jay Anderson

Physics & Astronomy Faculty Publications

Using Hubble Space Telescope (HST) imaging, we searched for candidate magnetically active binary (AB) counterparts to low-luminosity Chandra X-ray sources in the globular cluster 47 Tucanae (NGC 104). We consider the catalog of 300 Chandra sources within the half-light radius of 47 Tuc compiled by C. O. Heinke et al. (2005). The 10-band HST imaging used to identify these sources spans the range from 275–658 nm. We generated a cross-matched catalog of objects detected by photometric reductions. We used this catalog to search for counterparts within Chandra source error circles using color–magnitude diagrams and a color–color plot. We …

Development And Evaluation Of An Mpas-Based Urban Modeling System For The New York City Region, Marco S. Yu, Kristian Rice

Publications and Research

Complex geometry and varying land cover make urban environments difficult to simulate accurately. This research presents a performance evaluation of the newly developed urban modeling system in the Model for Prediction Across Scales (MPAS). Weather over the New York metropolitan area during June 2018 was simulated using supercomputing resources and the new urban modeling system. Modeled 2-m air temperature was compared with historical hourly observations from five NYC Mesonet stations, one located in each of the five boroughs. Comparison showed that the model follows the timing of warming and cooling well, but overestimates the magnitude of temperature swings at peaks …

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 …

Software Engineering In The Age Of Coding Agents: Failure Modes And Rejection Patterns, Mahd Mohd Hindi

Thesis/ Dissertation Defenses

This thesis investigates the real-world behavior of LLM-driven coding agents that generate code changes and submit pull requests (PRs) to public software repositories. As these tools evolve from autocomplete-style assistants into more autonomous agents, their contributions increasingly interact with socio-technical review processes (human reviewers, bots, CI/CD gates, and project norms). This thesis focuses on understanding why agent-generated PRs are accepted or rejected and what these outcomes reveal about current agent limitations in practical development workflows. The main objective of this thesis is to systematically characterize rejection patterns and failure modes of agent-generated pull requests in real repositories. Specifically, the thesis …