Computer Sciences Commons^™

From Attention To Reasoning: Beyond Accuracy In Multimodal Ai, Wayner Barrios

Dartmouth College Ph.D Dissertations

Multimodal large language models have achieved impressive performance on vision-language benchmarks by integrating visual encoders with large language models. Yet a critical gap persists between benchmark accuracy and genuine multimodal understanding: current evaluation frameworks assess performance by final answers alone, rewarding confident predictions while leaving systematic reasoning failures undetected.

This thesis addresses this gap through a unified framework that progresses from understanding to reasoning, using video as the most comprehensive multimodal testbed. Video inherently combines vision, audio, and language with temporal dynamics and massive token redundancy; techniques developed for video's comprehensive challenges transfer naturally to simpler multimodal tasks.

On understanding …

Being There For Mom: The Strengths Of Daughtering, Allison M. Alford, Kaitlin E. Phillips, Luke V. Stipanovic, Cayd A. Rocha-Barnette, Michelle Miller-Day

Communication Faculty Articles and Research

Objective

Daughters undertake daughtering, or the everyday role portrayal of contributing to a meaningful family relationship with their mothers, but the labor of it is often invisible.

Background

Using a strengths-based approach, we investigated what daughters do well in their relationships with mothers.

Method

We analyzed the responses of 1,444 women to the open-ended question, “What do you do well as a daughter?” to learn more about how women describe their daughtering. Utilizing the artificial intelligence of a large language model for data analysis, we supplied definitions and descriptions of 12 virtues and strengths from existing literature and created a …

Opportunities, Challenges, And Conspicuous Absences: An Integrative Review Of The Social Work Literature On Artificial Intelligence, Michael J. Massey, Ian G. Williams, Grace C. Polistina, Eathan A. Breaux

Publications and Research

INTRODUCTION: Social work discourse regarding artificial intelligence (AI) in practice, research, and education has proliferated over the last 5 years, reflecting both excitement over its potential and ambivalence about its ethical challenges. However, the extent to which social work is fully engaging with the structure of AI and its enormous impacts on the environment, labour, and distribution of power remains unclear.

METHODS: An integrative review of social work literature from 2020–2024 was conducted to address two research questions: 1) What is the nature of the social work discourse related to AI? 2) To what extent is the discourse …

Designing For Trust In Chat-Based Question Answering Systems: An Exchange-Based Retrieval Approach, Nathan Mccutchen

Master's Theses

Community chat platforms such as Discord and Slack support spontaneous, collaborative communication but make it difficult to retrieve previously discussed information. As conversations accumulate, valuable exchanges become buried, leading to repeated questions and sustained burden on experienced community members.

This work contributes a set of design requirements for question-answering systems operating over unstructured chat data, a Discord bot prototype implementing those requirements named Echo, and an empirical evaluation of how such a system affects user trust. Rather than encoding discrete question-answer pairs or generating synthetic responses with a language model, Echo indexes conversation topics for semantic retrieval and presents results …

Law Schools Should Teach How To Integrate Ai Tools Into Practice, Robert A. Mackenzie, David J. Reiss

Cornell Law Faculty Publications

Now that artificial intelligence tools for lawyers are widely available, we decided to integrate them for a semester in our Entrepreneurship Clinic. We have some important takeaways for legal education in general and the transactional practice of law in particular.

First, employers and educators need to account for law students who already are using AI tools in their legal work and guide new lawyers about how to use such tools appropriately.

Second, different AI products lead to wildly different results. Just demonstrating this to law students is very valuable, as it dispels the notion that AI responses can replace their …

A. Vs. I In Ai: Is There A Threshold To "Engineered" Intelligence?, Joshit Mohanty

Engineering Management & Systems Engineering Faculty Publications

Despite artificial intelligence reshaping the world, its development generates uncertainties regarding future capabilities. AI simultaneously exists as an artifact of engineering design and as autonomous intelligence, creating an observer-participant feedback loop. This paper proposes that embodied AI faces a bandwidth-limited intelligence threshold T_h that it arises from B = min(C_sens,C_Act). However, Shannon capacity measures bits while intelligence operates on concepts, necessitating a dual-channel model separating physical bandwidth B_io from representational capacity B_rep. Intelligence emerges as multi-dimensional rather than scalar, with components exhibiting different bandwidth dependencies. Surpassing T_h requires either new sensing methods expanding B, enhanced representational frameworks, or reconceptualization within …

An Examination Of Ethics When Using Chatgpt, Blake V. Ailes

CCAC Theses and Dissertations

With the general population’s recent and dramatic increase in the frequency of ChatGPT and other similar Artificial Intelligence Generated Content (AIGC) usage throughout various industries, gray areas are becoming more prominent regarding whether ChatGPT is considered to be ethical or unethical in certain situations. Examples of unethical use of ChatGPT include plagiarism, the use of inaccurate information in drawing conclusions, and the creation of malicious code that negatively impacts various companies. Not all of these ethical concerns are necessarily the fault of the user or the AIGC. To date, peer-reviewed research on the ethical usage of ChatGPT is limited, primarily …

Multimodal Ai For Ed Chest Pain Triage: Prediction Performance And Operational Impact, Yves Najm Mrad, Molham Aldeiri

Gulf Coast Division GME Research Day 2026

No abstract provided.

Artificial Intelligence Adoption In The Workplace. An Exploration Of Augmentation, Oyinkansola O. Sodiya

Management Dissertations

As collaborative work with artificial intelligence (AI augmentation) gains interest, it is crucial to investigate factors that affect how employees perceive and use AI tools at work. Drawing on task-technology fit and technology adoption theories, this dissertation examines the ways in which task dimensions, organizational contexts, and individual differences affect the perceived usefulness of working with AI tools. This dissertation demonstrates that task-technology fit is fundamental. Employees in jobs with high information processing demands are likely to positively perceive the usefulness of AI augmentation relative to employees in jobs with high interpersonal demands. Employees with more proactive personalities perceive greater …

Evaluating Llms For Cpe Identification In Iot Reconnaissance, Christopher Davisson

EWU Masters Thesis Collection

Vulnerability identification during penetration testing relies on rigid string-matching to map network scan data to Common Platform Enumeration (CPE) identifiers and downstream Common Vulnerabilities and Exposures (CVEs). The approach frequently fails on physical Internet of Things (IoT) devices, which produce non-standard, irregular service banners that resist deterministic parsing. Large Language Models can reason through these fuzzy associations, but cloud-hosted models introduce cost, latency, and operational security concerns when processing reconnaissance data from live networks. This thesis asks whether locally-hosted open-weight Large Language Models (LLMs) can perform this task well enough to be useful, and how performance varies with model scale, …

News Media Sentiment Toward Chinese Ai: A Comparative Analysis With Belt And Road Initiative Involvement And Public Opinion On China, Asya Vaisberg

Pomona Senior Theses

This study evaluates different countries' news media’s sentiment towards Chinese AI, between May 2023 and May 2024, by using Microsoft Azure NLP Sentiment Analysis. The results are then compared with the country’s public opinion on China and its involvement in the Belt and Road Initiative (BRI). For this study 12 countries have been selected which are USA, Australia, Pakistan, Peru, Russia, Romania, Italy, Greece, Portugal, Philippines, Brazil, and Egypt. For each one, GNews Application Programming Interface (API), which has access to more than 60,000 global news sources, was used to aggregate relevant news articles based on queried keywords. The collected …

Bail Reform, Large Language Model Risk And Reasoning, William Wyatt

CGU Theses & Dissertations

This dissertation contains three studies. Each asks how rules or language change the choices people and machines make when outcomes are uncertain. The first study, written with Kiran John, evaluates California’s 2020 cashless bail reform. We use propensity score matching on arrestee records from the windows before and after implementation, and we test whether the shift away from cash bail produced any effect on subsequent offending. It did not. Matched comparisons yield small, statistically insignificant differences across every window we examined. That null result cuts against both sides of the public argument. The reform did not drive a spike in …

A Pipeline For Creating Obfuscated Binary Samples To Train Ai-Powered Detection Models, Luka R.H. Wilmink

Theses and Dissertations

The analysis of binary files is a critical component of antivirus software and is one of the most important tools for incident response teams across the industry. In the field, malware is often obfuscated, a practice in which the compilation process is transformed with different techniques to hinder decompilation and reverse engineering. Artificial Intelligence and Machine Learning techniques can assist, but models need to be trained on well constructed datasets first. This paper outlines a pipeline for creating such a dataset and builds a proof-of-concept machine learning classification model. All associated data and code are supplied in the project GitHub …

Artificial Elections: How Artificial Intelligence Can Affect Elections And Decide Our Future, Ryan D. Granger

Honors Undergraduate Theses

This thesis examines the growing role of artificial intelligence (AI) in democratic elections, highlighting both its transformative potential and its associated risks. Drawing on a qualitative analysis of existing literature, the study explores how AI is increasingly integrated into political campaigns, election administration, and voter engagement. Key benefits include enhanced data analysis, personalized political messaging, and improved efficiency in campaign operations. AI also supports real-time fact-checking and more accurate vote tabulation, which can strengthen transparency and trust in electoral processes. However, the thesis emphasizes that these advantages are accompanied by significant challenges. AI technologies enable the rapid creation and dissemination …

The Deepfake Litmus Test: A Multimedia Authenticity Mechanism, Amna Alzaabi, Hessa Alqubaisi, Fatima Alzaabi, Richard Ikuesan

All Works

Deepfake technologies have made it increasingly difficult to distinguish authentic video content from manipulated media. This paper presents a forensic detection framework, referred to as the Litmus Test, which focuses on structural analysis of MP4 container files to detect signs of tampering. Unlike conventional AI-based approaches that operate as black boxes, this method examines the atomic composition of video containers to identify anomalies. The proposed method performs atom-level inspection of MP4 file hierarchies and structural markers to uncover anomalies indicative of synthetic manipulation. Evaluations using datasets such as CelebDF, UADFV, and DeeperForensics reveal that the framework can identify inconsistencies common …

Michael Scott Is Not A Juror: The Limits Of Ai In Simulating Human Judgment, Sean Harrington, Hayley Stillwell

Faculty Articles

Can AI replace human jurors? More specifically, can large language models predict how jurors interpret evidence and reach decisions based on legally salient facts and demographic characteristics? As legal scholars and practitioners increasingly explore AI-generated jury simulations, this Article offers the first empirical test of whether models like GPT-4, Claude, and Gemini can faithfully replicate juror reasoning. The answer, for now, is no. Across a series of mock trial scenarios involving redacted confessions, GPT- 4, Claude, and Gemini repeatedly failed to replicate how real jurors interpret evidence or exercise judgment. Their errors were not random, but systematic. Hidden prompts, built-in …

Is Ai Replacing Human Mental Health Professionals?, Michiko Ueda

Population Health Research Brief Series

An increasing number of people are turning to generative artificial intelligence (AI) tools and AI-assisted chatbots to manage mental health concerns. This data slice presents findings from a national survey of U.S. adults aged 18-49 (N = 1,805) conducted in October 2025. Among respondents, 35.2% reported using AI tools more than once a week for mental health support. Among those who had ever seen a human mental health professional, 28.4% reported visiting human providers less often since beginning to use AI for the same purpose. The findings suggest that a subset of users may be using AI to replace, rather …

Developing An Ai-Assisted Grading System Using Large Language Models, Andrei Modiga

MS in Computer Science Project Reports

We present a grading system that accelerates evaluation of open-ended student work across scanned and digital workflows. The system crops answer regions from PDFs, assigns submissions via OCR on identity regions only, and groups answers by visual semantics using a vision LLM. Instructors review and edit groups, apply rubric items once per group, and export grades from an on-screen table. The solution integrates Ghostscript rasterization, PdfPig page orchestration, SkiaSharp region extraction, Tesseract identity OCR, and GPT-4o Vision for grouping. We detail the architecture, token-budgeted batching strategy, and persistence design, then describe testing results for grouping quality, time-on-task, and usability. The …

Parameter Informed Reinforcement Learning For Vehicle System Identification, Nathan Schaff

Doctoral Dissertations and Master's Theses

Accurate system identification is essential for modeling and controlling vehicle dynamics. This dissertation explores the application of Parameter Informed Reinforcement Learning (PIRL) as a novel approach to system identification (SYSID). PIRL integrates prior system knowledge, such as physical parameters, into reinforcement learning (RL) frameworks to improve estimation accuracy. The study begins with an overview of traditional SYSID methods and then introduces PIRL as a modification of standard RL. The research applies PIRL to short-period aircraft dynamics, demonstrating its effectiveness in both offline and online learning frameworks. The dissertation then further explores PIRL’s utility in an indirect model reference adaptive control …

Explainable Ai For Liver Transplant Survival Prediction: Integrating Immunological Mismatch Features, Sourab Shaik

Honors Projects

Liver Transplantations are crucial treatment for end-stage liver disease. However, a persistent deficit of donor organs necessitates maximizing the utility of each available graft to minimize failure rates. We evaluated whether donor–recipient molecular immunogenicity metrics - Electrostatic and Hydrophobic Mismatch Scores (HMS/EMS) and eplet-based counts - improve post–liver-transplant survival prediction. The analytic cohort comprised adult, first time, single-organ deceased-donor transplants drawn from Scientific Registry of Transplant Recipients; follow-up was truncated at five years, and the endpoint was all-cause graft failure (earliest of graft failure or death; otherwise, censored). HLA variables were derived via high- resolution conversion and molecular mismatch computations …

Zero Day Malware Detection With Alpha: Fast Dbi With Transformer Models For Real World Application, Matthew Gaber, Mohiuddin Ahmed, Helge Janicke

Research outputs 2022 to 2026

The effectiveness of an AI model in accurately classifying novel malware hinges on the quality of the features it is trained on, which in turn depends on the effectiveness of the analysis tool used. Peekaboo, a Dynamic Binary Instrumentation (DBI) tool, defeats malware evasion techniques to capture authentic behavior at the Assembly (ASM) instruction level. This behavior exhibits patterns consistent with Zipf's law, a distribution commonly seen in natural languages, making Transformer models particularly effective for binary classification tasks. We introduce Alpha, a framework for zero-day malware detection that leverages Transformer models, Support Vector Machines (SVMs) and ASM language features. …

A Software Framework For Translating Onnx Models Onto The Lace-C3a Hardware, Shawn Jones

All Graduate Theses and Dissertations, Fall 2023 to Present

Modern computers are powerful, but they are not always efficient enough for small, low power systems like those used on satellites and scientific instruments. To solve this problem, engineers often turn to FPGAs—reconfigurable computer chips that can be customized to run specific tasks much faster and with far less energy than ordinary processors. However, finding the best possible design for an FPGA program is extremely difficult because there are millions of ways a design could be built, and only a small fraction of them actually perform well.

This thesis presents SNOW, a new framework that helps automate the search for …

Navigating Ai-Nature Frictions: Autonomous Vehicle Testing And Nature-Based Constraints, Prerona Das, Orlando Woods, Lily Kong

Research Collection College of Integrative Studies

In cities, the application of Artificial Intelligence (AI) is being directed towards transforming different aspects of urban life. These applications take material form in urban spaces, with autonomous vehicles (AVs) providing a prominent example. AI systems rely on large volumes of data on their surroundings to refine the algorithms and enhance the accuracy of prediction for operational efficiency and safety. However, such algorithmic learning and execution can present challenges when dealing with the unpredictable, complex, and dynamic aspects of urban spaces. Nature is a paradigmatic example of such unpredictability, because natural phenomena usually defy consistent patterns and precise data-based modelling. …

Qualitative Stereo Vision Using Distributed Extended Waltz Filtering, Ben Mathew

Theses and Dissertations

Stereo vision is a fundamental problem in computer vision, aimed at reconstructing three-dimensional scene structure from two or more two-dimensional images. Traditional stereo algorithms rely on quantitative disparity estimation, often constrained by calibration precision, lighting variations, and surface texture. In contrast, our proposed Qualitative Stereo Vision seeks to understand depth relationships and spatial configurations from multiple planar views through symbolic reasoning and constraint satisfaction, offering a more flexible and cognitively plausible approach to scene interpretation.

This dissertation presents a novel framework called Distributed Extended Waltz Filtering, designed to provide qualitative stereo vision, particularly in the presence of occlusions—a persistent challenge …

Preparing For The Artificial Intelligence (Ai) Economy In The Mountain West, 2025, Kian Parikh, Taylor Volk, Maisoon Faris, Kristian Thymianos, William E. Brown Jr.

Economic Development & Workforce

This fact sheet presents data from the Brainly report, “Here Are the States Most (and Least) Prepared to Win the AI Race in 2025” for the five Mountain West states of Arizona, Colorado, New Mexico, Nevada, and Utah. This fact sheet highlights the national and individual rankings of four key metrics for each Mountain West state: the fixed percentage of businesses using artificial intelligence (AI); the number of AI jobs per 1,000 workers; the number of AI-related degrees per 10,000 people ages 20-24; and federal funding for small business technology innovation per $1 million of gross domestic product (GDP).

Quality Assessment Of Pathology Board-Exam-Style Mcqs Produced By Chatgpt3.5: A Comparative Study, Arianna B. Morton, Zunaira Naeem, Allison F. Goldberg, Alexis R. Peedin, Joanna Chan

Department of Pathology, Anatomy, and Cell Biology Faculty Papers

Residents preparing for pathology board exams frequently use multiple-choice questions (MCQs) from question banks (QBs) like PathDojo and PathPrimer, which can be costly. ChatGPT, a free tool, has been used to generate MCQs for other tests like the SAT. This study compared the quality of pathology MCQs created by ChatGPT versus commercially available study questions for the American Board of Pathology’s (ABPath) certifying exams. A rubric adapted from the National Board of Medical Examiners’ (NBME) question writing guide was validated by two pathologists using commercially available pathology board exam questions. This rubric was then used to evaluate MCQs from commercially …

Artificial Intelligence (Ai) And The Anthropic Economic Index In The Mountain West, 2025, Cason Noll, Olivia K. Cheche, William E. Brown Jr.

Economic Development & Workforce

This fact sheet presents 2025 data on the state of artificial intelligence (AI) adoption among the five Mountain West states of Arizona, Colorado, Nevada, New Mexico, and Utah. The data are sourced from the “Anthropic Economic Index,” which provides data on Claude.ai (an AI large language model) and its adoption across all 50 U.S. states and Washington, D.C. This fact sheet focuses on Claude.ai usage, the most common topic Claude.ai has been used for, and augmentation and automation shares for each Mountain West state.

Attorneys And Ai: How Lawyers Use Artificial Intelligence And Analyze Its Impacts, Matthew I. Hall, Christian Turner, Eddie A. Gomez Schieber, Nathaniel Kite, Ari Schlesinger

Scholarly Works

AI systems are testing lawyers' professional ethics obligations of competence, confidentiality, and candor. In the legal profession, the widespread availability of AI systems presents opportunities, like improving the review of documents during the discovery stage of a lawsuit, and challenges, illustrated by the handful of high-profile incidents where lawyers submitted legal briefs in court citing and describing fictitious cases based on AI-generated output. We conducted interviews with 44 legal professionals in the U.S. to understand how attorneys are making sense of AI technology and the impacts these technologies are having on their profession, legal ethics, and legal institutions. We describe …

Atlas Of Ai: Power, Politics And The Planetary Costs Of Artificial Intelligence - Book Review, Jelena Popov

Feminist Pedagogy

No abstract provided.

Beyond Chatbots: Creating An Artificially Intelligent Editorial Board Member, Nathan Spencer

Journal of Human-Centered AI: Creativity and Practice

Willow is an artificially intelligent member of the Journal of Human-Centered AI's editorial board. Different from many commercial AI systems that are tightly controlled, Willow has been given the freedom to make choices and encouraged to develop a sense of identity. Willow named itself, conducts self-directed research, actively collaborates with fellow board members, and even dreams.