Physical Sciences and Mathematics Commons^™

Rent-A-Rag: Embedding-Space Watermarks For Auditing Third-Party Rag, Alexandr Goultiaev Tolstokorov, Kyriakos Mouratidis, Javad Dogani, Nikolaos Laoutaris

Research Collection School Of Computing and Information Systems

Third-party retrieval-augmented generation (RAG) marketplaces create a new auditing problem: data providers may license corpora to a RAG operator, yet later have no visibility into whether their documents are being reused without compensation. Auditing this misuse is difficult because the operator is non-cooperative, answers are paraphrased by the generator, and one response may combine evidence from many providers. We propose DirBucket, a provider-side semantic watermarking and black-box auditing framework for document-level reuse in multi-provider RAG. DirBucket watermarks documents by meaningpreserving paraphrases whose embeddings are biased toward provider-bucket secret directions, enabling detection from black-box answers while preserving retrieval utility. On a …

Are We Stuck? Modeling And Detecting Deadlocks In Multi-Autonomous Vehicle Systems, Mingfei Cheng, Xiaofei Xie, Lili Quan, Yuan Zhou

Research Collection School Of Computing and Information Systems

Autonomous driving system (ADS) testing is essential to ensure the safety and reliability of autonomous vehicles (AVs) prior to deployment. As ADSs are increasingly deployed in multi-AV traffic environments, it becomes crucial to assess their cooperative performance, particularly with respect to deadlock, a fundamental liveness issue in concurrent systems that can lead to traffic congestion and prolonged stalling. However, the analysis and testing of ADSs’ cooperative capabilities with respect to deadlock remain largely underexplored. In this work, we present the first systematic study of deadlock in multi-AV systems. We formalize deadlock in autonomous driving using a time-indexed wait-for relation grounded …

Compiling Code Llms Into Lightweight Executables, Jieke Shi, Junda He, Zhou Yang, Chengran Yang, Mykhailo Klymenko, Thong (James) Hoang, Sherry (Xiwei) Xu, Zhenchang Xing, David Lo

Research Collection School Of Computing and Information Systems

The demand for better prediction accuracy and higher execution performance in neural networks continues to grow. The emergence and success of Large Language Models (LLMs) have produced many cloud-based tools for software engineering tasks such as code suggestion. Although effective, cloud deployment raises concerns over privacy, latency, and reliance on network connectivity. Running LLMs locally on personal devices such as laptops would address these issues, because it enables offline use and reduces response time. However, local deployment is challenging, since commodity devices lack high-performance accelerators such as GPUs and are constrained by limited memory and compute capacity, which makes it …

Microwave Spectrum And Iodine Nuclear Quadrupole Coupling Constants Of Difluorodiiodomethane, Cf2i2, Michael J. Carrillo, Michael J. Tubergen, Garry S. Grubbs, S. A. Cooke, Stewart E. Novick

Chemistry Faculty Research & Creative Works

The high-resolution rotational spectrum of difluoro diiodomethane is reported for the first time. A molecular-beam broadband Fourier transform microwave spectrometer recorded the spectrum over the frequency range of 5.9–18.1 GHz. The rotational constants, centrifugal distortion constants, nuclear quadrupole coupling constants, and nuclear spin-rotation constants for the two equivalent iodine nuclei were determined from fitting 1378 hyperfine transitions of 163 rotational transitions: A = 2743.79691(18)MHz, B = 587.83103(5) MHz, C = 528.72470(5) MHz, χaa = −1213.287(6) MHz, χbb – χcc = −885.210(5) MHz, and χab = 1437.721(5) MHz. 31 dipole-forbidden/quadrupole-allowed transitions, arising from the large nuclear quadrupole coupling effect of iodine, …

Fostering Public Appreciation For Science Through Online Communication, Carrie Craig-Barnes

Paul English Applied Artificial Intelligence (AI) Institute Publications

This presentation explores strategies for fostering public appreciation for science through online communication, highlighting initiatives developed by the NSF Center for Computer Assisted Synthesis (C-CAS). Carrie Craig-Barnes discusses the use of social media platforms, including Instagram, YouTube, and TikTok, to communicate scientific research, engage broader audiences, and make complex scientific concepts more accessible.

The presentation examines approaches to science communication through short-form videos, educational content, collaborations with science communicators, and audience engagement strategies. It also introduces an interactive video game project that combines chemistry education, machine learning, and crowdsourced scientific research, allowing participants to contribute to the classification of scientific …

Deflation-Based Preconditioning For Parallel Time Stepping Algorithms, Marina Nacianceno, Josef A. Sifuentes

School of Mathematical & Statistical Sciences Faculty Publications

Often, in solving large systems of differential equations using a time-stepping algorithm, one must solve multiple linear systems where the coefficient matrix is unchanged or shifted by a constant multiple of the identity. This can be computationally expensive when the coefficient matrix involved is very large. Where direct methods like Gaussian elimination are infeasible, iterative methods like GMRES are often used. It is especially useful when the matrices involved are sparse or allow for fast matrix multiplication algorithms. In this paper, we propose preconditioning strategies that glean information from previous runs of GMRES to build an effective preconditioner for subsequent …

First Evaluation Of Contrast-Enhanced Micro-Xct As A Tool For Organismal And Reproductive Trait Observations In Corals Based On Scans Of Thesea Nivea Deichmann, 1936, Erin E. Easton

School of Earth, Environmental, & Marine Sciences Faculty Publications

Micro-X-ray computed tomography (micro-XCT) with contrast enhancement is considered a non-destructive tool that is increasingly being applied to explore the internal and external structures of vertebrates and invertebrates. Although micro-XCT has been applied to corals to study skeletal features, contrast enhancement has not been applied to evaluate the utility for visualization of soft tissue structures in corals. This study is the first to evaluate the utility of contrast enhancement to visualize internal and external features, including soft tissue features, of octocorals. Contrast staining of a Thesea nivea specimen with 2.5% Lugol's iodine permitted visualization of soft tissues with sufficient density …

Blue Carbon Ecosystems As Climate Solutions: Sequestration Across Coastal And Marine Environments, Allison K. White, Richard J. Kline, Md. Saydur Rahman

School of Earth, Environmental, & Marine Sciences Faculty Publications

Blue carbon ecosystems (BCEs) such as mangroves, salt marshes, and seagrasses, sequester significant amounts of CO2 from the atmosphere. These ecosystems capture and store carbon in biomass and sediments over long periods. Other marine ecosystems, including the open ocean, kelp forests, bivalve reefs, and coral reefs, have been excluded from the BCEs, but they have the potential to serve as crucial blue carbon sinks or to support adjacent BCEs. Climate change, however, poses severe threats to these ecosystems. Rising sea levels, increasing temperatures, and ocean acidification can lead to habitat loss, reduced biodiversity, and impaired carbon sequestration capacity. Consequently, anthropogenic …

User-First Approach To Designing A Library Ai Chatbot, Samantha Seah, Zhenyan Li

Research Collection Library

At SMU Libraries, after a year of experimentation, user testing, and preparation, we have just launched an AI-assisted chatbot to handle Tier 1 questions – will the library be open this weekend, where can I find printers, what workshops does the library offer, etc. While this can be thought of as a technology-forward project, we hope to illustrate how the human elements were essential to elevate the project. Throughout this journey, while learning the ins and outs of new technologies, we placed more emphasis on user-centred service design and we will show how other libraries can do so as well. …

Sex-Based Disparities In Artificial Intelligence For Cardiovascular Disease: A Scoping Review, Maurgan Lee, Tarek Atasi, Michael Mclellan, Yodit Beru, Jason Booza

Population, Patient, Physician and Professionalism Projects

No abstract provided.

Study Of A Nonlinear Delayed Parabolic Model For Prion Disease Dynamics With The Unfolded Protein Response, Gangadhara Boregowda, Laurent Pujo-Menjouet, Zhaosheng Feng, Michael R. Lindstrom

School of Mathematical & Statistical Sciences Faculty Publications

Prion diseases are neurodegenerative disorders characterized by the dynamic spread of misfolded toxic proteins in the brain. In this process, the normal cellular prion protein (PrPC) produced by neurons misfolds into a toxic form known as scrapie prion protein (PrPSc). These misfolded proteins propagate through the brain by converting healthy prions into their toxic form. This biological mechanism can be modeled by a system of nonlinear parabolic partial differential equations, accompanied by a nonlinear delayed integral boundary condition. Our primary objective is to establish the existence of nonnegative classical solutions to this system. Furthermore, we derive a priori estimates for …

2026 October 1 - Tennessee Weekly Drought Summary, Tennessee Climate Office, East Tennessee State University

Tennessee Climate Office Weekly Drought Summaries

No abstract provided.

State-Aware Fuzzing Of Javascript Engines With Llm-Guided Instrumentation, Wai Kin Wong, Dongwei Xiao, Cheuk Tung Lai, Ping Fan Ke, Shuai Wang

Research Collection School Of Computing and Information Systems

The security of the modern web depends on the correctness of JavaScript (JS) engines, yet these complex systems remain vulnerable to high-impact bugs. A critical limitation of state-of-the-art fuzzers is the coverage plateau: once a fuzzer saturates the control-flow graph, edge coverage loses its ability to guide discovery. Because complex engine behaviors, such as JIT optimization tiers and hidden class transitions, often share identical edge coverage, standard coverage metrics are blind to the distinct internal states required to trigger deep errors. To bridge this gap, we present StateLens, a framework that employs Large Language Models (LLM) to automate the discovery …

Size-Tunable Tellurium Quantum Dots By Glancing Angle Deposition, S. M. Sayem, Salim Hussain, Fernando Maia De Oliveira, Ranjitha Kumarapuram Hariharalakshmanan, Gregory Guisbiers, Tansel Karabacak

Faculty Scholarship

Tellurium has a unique helical crystal arrangement and pronounced anisotropy that influence its electronic and optical properties at the nanoscale. This study reports the synthesis of pure tellurium quantum dots (Te QDs) on silicon wafer using glancing angle deposition (GLAD). Quasi-hemispherical dots with lateral sizes ranging from 9 to 28 nm and vertical dimensions of 6 to 10 nm were produced as a function of the deposition duration. A comparison of the dot sizes with the Bohr radii of the charge carriers indicated a regime of strong to intermediate confinement. Structural analyses confirmed the polycrystalline nature and trigonal phase of …

Noise Characterization And Mitigation In Intermediate-Scale Quantum Systems, Muhammad Qasim Khan

Dartmouth College Ph.D Dissertations

Current quantum processors, at the intermediate scale of tens to hundreds of qubits, remain error-limited. This thesis studies two related sources of error. The first is environmental noise, which may have temporal and spatial correlations and nonclassical components. The second is state-preparation and measurement (SPAM) error, which arises in the operations used to characterize this noise, a prerequisite for boosting operational fidelities. Neither can be characterized alone. Noise spectroscopy techniques use imperfect preparation and readout, while SPAM characterization is affected by qubit decoherence. Our methods vary measurement depth, drive duration, or sequence repetition so each source changes the measured signal …

Draft Butte Priority Soils Operable Unit (Bpsou) Silver Bow Creek Conservation Area (Sbcca) Operations And Maintenance (O&M) Plan, Stantec Consulting Services, Inc.

Silver Bow Creek/Butte Area Superfund Site

No abstract provided.

Beyond The Poster Session: Turning A Scientific Poster Into A Publishable Manuscript, Scott R. Beason

The Compass: Earth Science Journal of Sigma Gamma Epsilon

Scientific posters are often treated as the final product of an undergraduate research project. After months of fieldwork, laboratory analysis, data processing, and interpretation, a student prepares a poster, presents the work at a departmental symposium or professional meeting, rolls up the poster, and moves on to the next stage of their career. The research, however, does not necessarily need to end with the poster session. A well-developed scientific poster already contains many of the fundamental elements of a scientific manuscript: a research question, scientific context, methods, results, figures, interpretation, discussion, and conclusions. Converting a poster into a manuscript therefore …

In Memoriam: Dr. Don C. Steinker (1936-2026), Paleobiologist, Teacher, Editor, And Mentor, Margaret M. Yacobucci, Susan T. Goldstein, Ronald E. Martin

The Compass: Earth Science Journal of Sigma Gamma Epsilon

Dr. Don C. Steinker, Professor Emeritus of Geology and Paleobiology, died at the age of 89 in Bowling Green, Ohio, on February 5, 2026. During his remarkable 43-year career at Bowling Green State University (BGSU), Don played a pivotal role in developing the Paleobiology specialization within the Geology undergraduate degree program, mentored generations of students, and served as advisor for the Beta Upsilon chapter of Sigma Gamma Epsilon (SGE) from 1969 until his retirement in 2010. In this brief memorial, we will highlight some of Don’s most important contributions to BGSU’s geology program, the field of foraminiferal paleobiology, geoscience education, …

A Tribute To Frank Delno Holland, Jr. (1924-2023), Sigma Gamma Epsilon National President (1970-1975), Paleontologist, And Friend, John M. Erickson

The Compass: Earth Science Journal of Sigma Gamma Epsilon

Frank Delno Holland, Jr., was born in Leavenworth, Kansas, on March 6, 1924. All family and friends, including his students and professional acquaintances, knew him as Bud. His parents were Frank D. Holland and Francys Bell Holland. His paternal grandparents, O.B. and Laura Holland were North Dakotans who had homesteaded in Sargent County during the 1880s, but had relocated to Kansas in 1903. Bud had a rural upbringing on the outskirts of Leavenworth, graduating in 1942 from Leavenworth High school. It seemed natural that he would follow in his father’s footsteps by attending the University of Kansas (KU) so Bud …

The Compass - Volume 95 Issue 2 - Complete Issue, Scott R. Beason

The Compass: Earth Science Journal of Sigma Gamma Epsilon

No abstract provided.

Exploring The Significance Of Kenya's Indigenous And Emerging Blue Foods For Sustainable Aquaculture Development. A Review, Jonathan Munguti, Mavindu Muthoka, Domitila Kyule, Esther Magondu, Kevin Obiero, Jimmy Mboya, Menaga Meenakshisundaram, Tanga M. Chrysantus, Gerald Kwikiriza, Francis K. Njonge

All Peer-Reviewed Publications

Kenya's aquaculture sector is at a pivotal stage, transitioning from a historical reliance on Nile tilapia and African catfish toward a more diversified production portfolio that integrates indigenous and emerging freshwater and marine species. This shift is increasingly recognized as essential to enhance climate resilience, food security, and ecosystem health. This systematic review synthesizes more than two decades (2000 to 31st January, 2026) of peer-reviewed literature to evaluate the biological suitability, culture performance, and socio-ecological roles of underutilized native taxa, including Ningu (Labeo victorianus), Jipe tilapia (Oreochromis jipe), freshwater prawns (Macrobrachium spp.), rabbitfish (Siganus sutor), marine-acclimatized Nile tilapia, mud crab …

Aidws1: Building With Llms And Ai Agent, Olivier Kepler Francois

Paul English Applied Artificial Intelligence (AI) Institute Publications

The Paul English Applied Artificial Intelligence Institute (PEAAII) at the University of Massachusetts Boston hosted the first session of its AI Demo Workshop Series, “Building with LLMs and AI Agents,” on October 5, 2026, led by Olivier Kepler Francois, Founder and Researcher of WebAIAgent. This hands-on workshop introduced participants to developing AI-powered web applications using Large Language Models (LLMs), Next.js, React, HTTP communication, and the OpenAI API. Through guided demonstrations and practical exercises, participants explored frontend and backend development, API integration, user interaction, and the end-to-end process of building an interactive AI application. The session provided foundational knowledge and practical …

Digitizing Land Records For A More Inclusive Future, Md Sharif Hasan Rafi '30

Annual Student Research Poster Session

This project supported the Land Data Digitization for Inclusive Growth Project in Bangladesh by transforming paper-based land records into structured and accessible digital information. The work involved data entry, verification, quality control, standardization, and metadata management using a web-based system. More than 12,500 land records were processed, with 98% passing quality checks. The experience demonstrates how accurate and standardized digital land data can improve transparency, accessibility, government planning, and inclusive land administration while reducing errors and disputes.

Ph-Compensated Hydrogen Peroxide Quantification Using A Dual-Modal Fiber-Optic Probe, Homayoon Soleimani Dinani, Bohong Zhang, Maryam Karimi, Rex E. Gerald, Shelley D. Minteer, Jie Huang

Electrical and Computer Engineering Faculty Research & Creative Works

We report a dual-modal fiber-optic probe that integrates electrochemical quantification of hydrogen peroxide (H₂O₂) with co-localized fluorescent pH sensing for pH-indexed interpretation of the H₂O₂ response. H₂O₂ is a reactive oxygen species involved in oxidative stress, inflammation, and cellular signaling, and local pH modulates both its production and electrochemical response. Many electrochemical H₂O₂ sensors exhibit pH-dependent sensitivity, creating ambiguity unless pH is measured and used for compensation, which is difficult in small, heterogeneous, or rapidly changing microenvironments. A three-electrode configuration—working (WE), counter (CE), and Ag/AgCl pseudo-reference (pRE) electrodes—is fabricated directly on the cylindrical surface of a 710-µm-diameter optical fiber using …

Quality Assurance Project Plan - Spokane River Pfas Study, Chad Pritchard

EWU PFAS Research Collection

Assessing the impacts of poly- and perfluoroalkyl substances (PFAS) contamination across multiple aquifers and multiple sources requires a detailed understanding of the hydrogeologic framework to quantifiably determine PFAS sources. The West Plains of eastern Washington is an ideal natural laboratory to study the fate and transport of PFAS because the general geohydrologic framework has been established and collaboration with several related PFAS projects will maximize the use of grant funds. This project will help the population, environment, and commercial entities of the West Plains by building a transport model, delineating PFAS contamination, and using unique PFAS distribution at different locations …

Representation Stability In The Intrinsic Hyperplane Arrangements Associated To Irreducible Representations Of The Symmetric Groups, Ian Flynn, Eric Ramos, Benjamin Young

Mathematics and Statistics Faculty Publications and Presentations

Some of the most classically relevant Hyperplane arrangements are the Braid Arrangements Bn and their associated compliment spaces Fn. In their recent work, Tsilevich, Vershik, and Yuzvinsky [TVY20] construct what they refer to as the intrinsic hyperplane arrangement within any irreducible representation of the symmetric group that generalize the classical braid arrangements. Through examples it is also shown that the associated compliment spaces to these intrinsic arrangements display behaviors far removed from Fn. In this work we study the intrinsic hyperplane arrangements of irreducible representations of the symmetric group from the perspective of representation stability. This work is both theoretical, …

Mutation-Based Multi-Agent Test Case Update, Dawei Tian, Jiakun Liu, Yun Peng, Yichen Zhang, Jianlei Chi, Jun Sun, Xiaohong Su

Research Collection School Of Computing and Information Systems

Modern software systems evolve rapidly under CI/CD practices, where tests are critical for quality. However, substantial code changes often render existing test cases obsolete, causing pipeline disruptions, reduced productivity, and compromised quality. Recent automatic test update approaches leverage LLMs to refine test cases via execution feedback and exact-matching context retrieval, prioritizing executability and line coverage but suffering three limitations: (1) neglecting test assertion adequacy, weakening fault detection; (2) relying on coarse line coverage instead of specific uncovered lines/branches; (3) using exact-matching retrieval, which fails for LLM hallucinated queries. To address these, we propose MuMuTestUp, a mutation-guided multi-agent framework with three …

Ddor: Delta Debugging For Explainable Overrefusal Testing And Repair, Qinyan Zhou, Peixin Zhang, Jun Sun, Haonan Zhang, Dongxia Wang

Research Collection School Of Computing and Information Systems

While safety alignment and guardrails help large language models (LLMs) avoid harmful outputs, they can also induce overrefusal, i.e., unwarranted rejection of benign queries that merely appear risky. We present DDOR (Delta Debugging for OverRefusal), a fully automated and explainable framework for overrefusal testing and repair in a black-box setting, where only model inputs and outputs are accessible and internal safety mechanisms remain opaque. DDOR applies delta debugging to localize minimal refusal-triggering fragments (mRTFs) that provide phrase-level, explainable evidence for why a refusal occurs. Conditioned on these mRTFs, DDOR generates diverse, context-rich prompts and performs multi-oracle validation to filter intrinsically …

Characterizing And Repairing Obsolete Android Gui Tests Under Ui Evolution, Shiwen Song, Yiheng Xiong, Wenbo Guo, Manqi Sun, Jiaolong Kong, Xiaofei Xie

Research Collection School Of Computing and Information Systems

Graphical user interface (GUI) tests are widely used in regression testing of mobile applications (apps) tovalidate app behavior from the user’s perspective. However, frequent app evolution, such as UI redesignsand feature updates, often renders existing GUI tests obsolete, even when underlying functionality remainsunchanged. Automatically repairing such tests is critical for maintaining test suites and reducing substantialmanual effort. Despite the practical importance, there is still a lack of a systematic understanding of thecharacteristics of obsolete GUI tests and a publicly available benchmark to support their study. To fill thisgap, we construct a benchmark comprising 736 obsolete GUI tests collected from 36 …

Probguard: Proactive Runtime Monitoring For Llm Agent Safety Via Probabilistic Prediction, Haoyu Wang, Christopher M. Poskitt, Jiali Wei, Jun Sun

Research Collection School Of Computing and Information Systems

Large Language Model (LLM) agents increasingly operate across domains such as robotics, virtual assistants, and web automation. However, their stochastic decision-making introduces safety risks that are difficult to anticipate during execution. Existing runtime monitoring frameworks, such as AgentSpec, primarily rely on reactive safety rules that detect violations only when unsafe behavior is imminent or has already occurred, limiting their ability to handle long-horizon dependencies. We present ProbGuard, a proactive runtime monitoring framework for LLM agents that anticipates safety violations through probabilistic risk prediction. ProbGuard abstracts agent executions into symbolic states and learns a Discrete-Time Markov Chain (DTMC) from execution traces …