Digital Commons Network^™

Proactive Mental Health Assistance Via Agentic Llm Chatbots With Retrieval-Augmented Generation, Shaira Wajiha

ATU Scholars Symposium

Mental health challenges such as anxiety and depression are widespread, yet often remain unaddressed due to stigma, financial barriers, limited access to professionals, or personal reluctance. To address this gap, we present an accessible AI-powered chatbot that provides preliminary conversational support with empathetic, context-aware responses. The chatbot is trained and evaluated on two publicly available counseling conversation datasets from Huggingface, enabling it to learn and maintain therapeutic dialogue patterns. We compare three advanced LLMs, such as Llama3.1, Mistral 7b, and Qwen3, evaluating them on relevance, empathy, conciseness, and contextual understanding, and all models demonstrate high response quality. Incorporation of a …

Developing Machine Learning Algorithms For Highly Imbalanced Neonatal Disorder Data, Ali Nawaz

Thesis/ Dissertation Defenses

Neonatal disorders such as low birth weight, very low birth weight, extremely low birth weight, preterm birth, and very preterm birth increase the likelihood of high neonatal morbidity or mortality and call for early identification. However, the rarity of occurrence of these conditions in the clinical datasets has resulted in a severe class imbalance, raising questions about the application of binary classification models to them. Therefore, this thesis proposes a sequential methodological framework for neonatal disorder detection under different assumptions related to the availability of labels. Initially, binary classification experiments are conducted to analyze the behaviour of commonly used classification …

Tavern Table: A Peer-To-Peer Online Platform For Collaborative Dungeons & Dragons Campaigns, Dustin Needham, Kristen Avery, Aidan Martin, Seth Poindexter

ATU Scholars Symposium

Designed to address the lack of free, accessible, and feature-complete online Dungeons & Dragons gameplay platforms, Tavern Table gives users the ability to create or participate in Dungeons & Dragons campaigns online via peer-to-peer multiplayer. This platform is targeted primarily for two sets of users: Dungeon Masters (the game masters), who will be creating and hosting campaigns for players to participate in, and the players participating in said campaigns. We chose the Unity Real-Time Development Platform to develop Tavern Table as it was a free and effective platform that supported 2D game development as well as peer-to-peer multiplayer. To create …

Simple D&D: A Web-Based Character Creation System For Dungeons & Dragons, Brennan M. Bondio, Dalton N. Tate, Landon F. Boyd

ATU Scholars Symposium

Dungeons & Dragons, published by Wizards of the Coast, is a widely played tabletop role-playing game that requires players to create detailed characters governed by structured rule systems. Character creation involves managing interdependent attributes, calculations, and constraints that can be difficult for new players and time-consuming even for experienced participants. These complexities create a barrier to entry and reduce efficiency during gameplay preparation.

This project addresses that challenge through the development of Simple D&D, a web-based character creation system designed to streamline and automate rule-driven character configuration. The system aims to reduce manual calculation errors and setup time while maintaining …

Automatic Labeling Of Real-World Pmu Data: A Weakly Supervised Learning Approach, Yunchuan Liu

Research Days

This paper presents a weakly supervised learning framework for real-world event identification in transmission networks using phasor measurement unit (PMU) data. The growing integration of renewable energy sources has introduced greater variability in grid conditions, intensifying the need for accurate event detection. Although high-resolution PMU measurements enable event identification to be formulated as a classification problem, traditional supervised learning approaches are hindered by the scarcity of labeled data, and acquiring large-scale, high-quality labeled PMU datasets remains prohibitively expensive. To overcome this challenge, we propose an automated PMU data-labeling method that combines domain knowledge with machine learning techniques through the use …

Digitizing Transportation Operations At Safe Haven, Luke S. Garrett, William B. Turk, Clay A. Curtis, Aiden H. Behler

ATU Scholars Symposium

Safe Haven’s transportation department currently relies on a paper-based documentation process that requires physical transfer of records between buildings and repeated manual uploading of documents into storage systems. This workflow creates delays, redundant administrative tasks, and increased risk of misplaced or inconsistent records. Drivers, transportation coordinators, reviewers, and clients all interact with this process, making efficiency and data accuracy critical to daily operations.

This project develops a web-based transportation scheduling system designed to digitize documentation workflows and automate many of the repetitive tasks. The system replaces physical records with digital data management, reducing unnecessary manual handling and improving information accessibility …

Roomiq, Po Kya, Eh Paw, Nicole Holt, Riley Lindsay

ATU Scholars Symposium

1. RoomIQ is a smart room scheduling system designed to replace the current booking process for Corley Room 230 at Arkansas Tech University (ATU). The existing system presents challenges in efficiency, accessibility, and real-time coordination. Our goal is to deliver a user-friendly, real- time coordinated reservation solution that improves both functionality and overall user experience. The system will operate on an iPad Mini mounted outside the room, allowing users to instantly check availability and reserve the space on-site. This provides a convenient solution for immediate scheduling needs. In addition, a QR code displayed at the entrance will allow users to access …

Blazewatch: A Web-Based Management System For Volunteer Fire Departments, Michael J. Heinzen, Locke G. Weisler, Hunter S. King, Ryan N. Williams

ATU Scholars Symposium

BlazeWatch is a web application developed for rural volunteer fire departments. Volunteer fire departments often rely on manual processes and paper records systems to manage district fire dues, emergency call documentation, and equipment tracking. These inefficiencies can increase administrative workload and reduce efficiency. BlazeWatch is a secure full-stack web application designed to streamline administrative management for volunteer fire stations while providing a public-facing interface for general department information as well as a way for users to pay their fire dues. The system is built using an Angular frontend and a .NET backend with Identity-based authentication to ensure secure access control. …

Memorra: A Mobile-First Reminder Application For Personal And Social Event Management, Aaron Ngo, Caden Cash, Asher Wise

ATU Scholars Symposium

Memorra is a mobile-first reminder application designed to help individuals manage important dates and social commitments in a secure and reliable manner. Many users—including older adults, students, working professionals, and individuals experiencing cognitive overload—struggle to consistently remember birthdays, anniversaries, deadlines, and other significant events amid demanding schedules. To address this challenge, Memorra provides a centralized platform for managing reminders and organizing events with timely push notifications. The system architecture employs Flutter and Dart for cross-platform mobile development, supported by Firebase services—including Firestore, Authentication, Hosting, and Cloud Messaging—for backend functionality, real-time data management, and notification delivery. Secure user authentication, encrypted data …

Moneyup: A Predictive Financial Management System For College Students, Isaiah J. Adams, Malaya E. Wilburd, Dan V. Le, Joshua P. Golden

ATU Scholars Symposium

College students often lack accessible tools that combine real-time financial tracking, mobile accessibility, predictive analytics, and secure system design, leaving many without structured insight into their spending behavior. MoneyUP is a full-stack financial management platform developed to address these challenges through a secure, data-driven budgeting system deployed as both a web application and a cross-platform Flutter mobile application. The system integrates the Plaid API in its Sandbox environment to synchronize simulated banking data for secure testing without exposing live financial credentials. Transaction data is processed and stored using Supabase with a relational PostgreSQL database structured to enforce normalization, referential integrity, …

Llm-Based Stock Sentiment And Market Intelligence Platform, Joshua Thrower, Andrew Pinkerton, Ian Duggan, Wyatt Lester

ATU Scholars Symposium

Financial markets increasingly react to social media discourse, yet investors lack tools to translate this unstructured commentary into measurable indicators. Platforms such as YouTube host extensive discussions about publicly traded equities, but extracting reliable sentiment trends from high-volume, noisy comment streams remains technically challenging. This project develops a stock sentiment and market intelligence platform that transforms YouTube comment data into aggregated sentiment indicators aligned to specific equities. Comments are mapped to equities using ticker specific keyword identification combined with contextual filtering to reduce false associations from ambiguous or off-topic mentions. The system assigns numerical sentiment scores to individual comments and …

Aiw26s: Applied Llms, Chengjie Zheng

Paul English Applied Artificial Intelligence (AI) Institute Publications

This workshop introduces the concept of applied large language models (LLMs), focusing on how users can move from simple prompt-based interaction to building structured, repeatable AI-driven workflows. Participants explore how AI enables faster prototyping, lowers barriers to entry, and expands who can participate in building technology. Through a hands-on demonstration, attendees learn how to transform raw inputs into meaningful outputs such as summaries, key concepts, and actionable steps. The session emphasizes the importance of clear problem definition, iterative refinement, and critical evaluation when working with AI systems.

Fully Decentralized Hierarchical Federated Learning At The Edge With Post-Quantum Secure Communication, Tariq Qayyum

Thesis/ Dissertation Defenses

Federated learning (FL) enables collaborative model training without centralizing raw data, but deploying FL at scale in real edge environments remains challenging because iterative training and aggregation must operate over heterogeneous, resource-constrained, and often mobile devices with time-varying connectivity. Conventional hierarchical federated learning (HFL) partially mitigates communication cost by introducing fog/edge aggregation, yet many designs retain cloud-based global aggregation and cloud-centric coordination. This places wide-area network latency on the critical path of every training round, creates a single point of failure, and limits responsiveness as model sizes and federation scale grow. Moreover, moving coordination and aggregation closer to the edge …

Romance Scam Reporting And Support: Help-Seeking Timing, Trust, And Escalation, Ld Herrera

Research & Publications

Built on social engineering and identity deception, romance scams often create financial loss and distress. For victims, it can be difficult to know where to go, what information is needed, and what outcomes are realistic. This paper reports results from an anonymous survey of people who were targeted by or experienced a romance scam (completed surveys: n=386), focusing on (1) when and whether victims first reach out for help, (2) perceived difficulty and confidence in navigating support, (3) how trust relates to expectations of assistance, and (4) how loss severity relates to transfer-method complexity. When help was sought, it was …

Ai Exposure And The Future Of Work: Tasks, Skill Demand, And Education, Erik Vasilauskas, Michael Horrigan

External Papers and Reports

No abstract provided.

Ai Method For Classification Of Diagnosis Of Near-Infrared Breast Lesion Images, Kaiquan Chen, Fangyang Shen, Honggang Wang, Zhengchao Dong, Jizhong Xiao, Ming Ma, Afroza Aktar, Christopher Chow, Wenxiong Zhang

Publications and Research

In near-infrared optical breast lesion screening and diagnosis systems, high-speed four-dimensional scanners can dynamically acquire tens of thousands of lesion images within a five-minute period. Currently, manual computer annotation is required to generate standard samples from these scanned breast lesion images, a process that depends heavily on physicians with clinical expertise. On average, a single physician can annotate only approximately ten samples per working day. As a result, this process is time-consuming and labor-intensive, and the collected samples often suffer from low accuracy, large variability, and limited diagnostic reliability. Several AI-based annotation tools, such as QuPath, HALO AI™, and X-AnyLabeling, …

Department Portfolio Web App*, Phillip Suvacarov, Tommy Aitchison

Campus Research Month

Southern Adventist University’s School of Computing produces numerous course projects, capstones, and research papers each year, yet there is no centralized, public showcase for this work. Our system provides a structured submission workflow for current and former students, faculty approval to ensure academic quality, and moderated commenting and likes to encourage constructive engagement. We outline the content model, role-based access control, and review queue, and describe search, tagging, and media support (including PDFs, images, and code links). By making student work visible beyond the classroom, the portfolio supports recruitment, alumni relations, and employer outreach while strengthening the School’s scholarly community.

Improving User Experience And Functionality: The Redesign Of Sorora In React Native*, Katherine A. Arroyo, Oswin S. Shin

Campus Research Month

This project redesigned and extended Sorora, a safety-focused mobile application built in React Native. The original Android Minimum Viable Product (MVP) included location tracking, an SOS button, and a basic contact list, but the interface lacked clarity, and the feature set created friction during urgent situations. We improved the UI and UX using Nielsen's usability heuristics, reduced navigation complexity, and redesigned the emergency workflow for deliberate, fast interaction.

Redesigning A Fitness App Interface For Physiological Constrained Users*, Joseph E. Manzanillo

Campus Research Month

As fitness tracking converges with medical monitoring, inclusive design becomes a matter of health equity. This research utilizes a Polar Beat redesign to address exclusionary "sporty" aesthetics that can exclude 300 million colorblind users. Based in Human-Computer Interaction (HCI), the study implements WCAG AA standards and color-blind-verified filters to mitigate data loss during Situational Induced Impairment (SIID), when high-intensity exercise compromises cognitive and visual processing. By optimizing user journeys for Paralympic and geriatric archetypes, this work demonstrates that accessibility is the essential bridge transitioning mobile fitness apps into viable, inclusive instruments for clinical medical use.

A Strategic Roadmap For Assessing And Educating On Personal Cybersecurity Practices In Universities*, Ryan Lopez, Ysani Peña

Campus Research Month

Universities face a common cybersecurity threat: their own users. Although organizations may meet compliance standards and implement robust security infrastructures, the individual user remains the weakest link. This is particularly evident in higher education institutions, where both students and employees are frequent targets of cyber threats due to a lack of cybersecurity awareness. This paper proposes a strategic roadmap for assessing university student bodies and employee populations through cybersecurity domains that directly affect personal cyber hygiene awareness and practice.

Our proposed roadmap was validated in a U.S. university by using a domain-focused survey and simulated phishing campaigns. After the identification …

Sau Physical Activity Website (Paw)*, Lisbette Sanchez, Alberto Elizondo, Caleb Tenold, Evelyn Shtereva, Siegwart Mayr

Campus Research Month

Problem: Southern’s Physical Activity Website (PAW) used to track physical activity from students and faculty was no longer functional. Aside from unsupported API versions, there was also an issue with authorization and role assignments.

Solution: Southern’s Center for Innovation and Research in Computing (CIRC) adopted the assignment of restructuring a new web application to track physical activity. This project focuses on building a secure and scalable architecture that connects user devices, third-party fitness APIs, and a centralized database to support activity tracking, workout analysis, and fitness history. By combining modern web development practices with API integration and backend data processing, …

Towards Physics-Informed Neural Networks For Simulating Multiphase Geothermal Convection​*, Daniel C. Patton, Andrew Harrison Eno

Campus Research Month

Water and steam flow through porous rock, transferring heat via conduction and buoyancy-driven convection caused by density differences. Traditional numerical methods (finite-volume/finite-element) model this well but can become memory-intensive and unstable for long, high-detail simulations. This work demonstrates a Physics-Informed Neural Network (PINN) using a finite-difference approach within the NVIDIA PhysicsNeMo framework to simulate magma chambers in 2D. Tested on the Rio Pisco pluton in Peru, results are compared with the USGS HYDROTHERM model. PINNs learn from physical laws, offering accurate, flexible solutions with less data and development effort.

Mass-Imaging Computers Over A Network For Southern's Information Technology Office*, Zane C. Meyers, Nicolas R. Goslee

Campus Research Month

Our research was meant to save time for Southern's IT department by researching and documenting a way to mass image computers in batches at a time over the network. The poster includes a introduction to our project, the research and testing process, and the results and conclusions drawn from this.

Chat Component For Php Web Apps*, Linton W. Feitosa, Caleb N. Hoffman

Campus Research Month

Chat rooms are a common feature in much of today's software. While many web applications could benefit from them, developing individual chat implementations can be repetitive and time consuming. Furthermore, publicly available and integrable chat components are difficult to find.

We created a general-use chat component for PHP web applications using the Yii2 framework. Our component features chat rooms, asynchronous messaging, and contact management. It is widely applicable, customizable, documented, and can be easily extended by future developers.

The Cake Is A Lie: Hid Wireless Adapter*, Andrew J. Patton, Benjamin Chant

Campus Research Month

This project explores converting wired Human Interface Devices (HID) into wireless devices by creating an adapter. Devices without wireless chips or dongles are hindered when flexibility is required, creating electrical waste. Our solution consists of a Transmitter (TX) and Receiver (RX) device pair and is designed to wirelessly bridge USB input from an HID device to a target host.

Towards Smaller Artificial Neural Network Using Mean Compression*, Michael D. Burks, Matthew K. Chuhng

Campus Research Month

Artificial Neural Networks (ANNs) require substantial memory and computational resources, limiting their deployment on resource-constrained devices. Our contribution is a compression method using Mean Compression (MC) to reduce ANN size while preserving functionality and accuracy. MC consolidates connections with similar edge weights into meta-nodes with averaged values. Unlike traditional pruning that only removes connections among neurons, MC restructures networks by recomputing weights and creating meta-nodes. Additionally, unlike fixed pruning thresholds, MC uses flexible weight range patterns. Applied to multilayer perceptron (MLP), ANNs are made more accessible for deployment on constrained devices as proven in several experiments. Specifically, across five classification …

Code Visualizer: An Interactive Code Visualization Tool, Tadd Trumbull

Campus Research Month

Code Visualizer is a web-based, interactive algorithm visualization tool designed to help introductory computer science students develop a deeper understanding of array searching and sorting algorithms. Code Visualizer presents a step-by-step simulation environment built on a restricted Python subset, which allows students to observe array traversal, index manipulation, and algorithmic operations in real time. The tool features two learning modes: View Mode and Predict Mode. The underlying architecture utilizes a behavioral software design pattern called command pattern.

Spatial Future Ahead! Augmented Reality And Anticipated Life Consequences, Sergio Barta, Reto Felix, Chris Hinsch, Mahdokht Kalantari, Nina Krey, Philipp A. Rauschnabel

Marketing Faculty Publications

Purpose: This study explores how initial exposure to immersive spatial computing experiences using AR headsets generates lasting inspiration and shapes consumers expected long-term life consequences (i.e., enhancement of reality, perceived substitutability and social impact).

Design/methodology/approach: The study uses a time-lagged research design based on 148 first-time users of spatial computing devices (AR headsets). Respondents were interviewed once shortly after being exposed to AR and a few days later. Data is analyzed using partial least squares structural equation modeling (PLS-SEM).

Findings: Users' immediate “inspired-by” experiences predict increased “inspired-to” intentions days later. Such inspiration translates into anticipated consequences such as virtually customizing …

Adopting Artificial Intelligence: Cross-Sector Analysis Of Ai Adoption Risks, Brandon Saari, Yona Berger, Yanett Munoz, Alex Agnick, Paul Wagner, Robert J. Honomichl

Journal of Cybersecurity Education, Research and Practice

Artificial intelligence (AI) is rapidly being adopted across public and private sectors. This offers significant gains in efficiency, decision making, and access to information. At the same time, AI introduces complex risks related to cybersecurity, privacy, bias, transparency, accountability, and equity that existing governance and security frameworks do not fully address. This paper presents a cross-sector literature review and comparative analysis of AI adoption risks and mitigation strategies across four critical domains: the federal government, libraries, K–12 education, and healthcare. Drawing on peer-reviewed research, institutional frameworks, and policy guidance, the study identifies sector-specific challenges alongside shared systemic gaps, including insufficient …

Innovations And Applications Of Virtual Private Networks And Sustainable Security In Society 5.0 Libraries, Stella Chinnaya Nduka Dr., Adeyinka Tella Prof, Petros Dlamini Dr

Journal of Cybersecurity Education, Research and Practice

In order to improve digital resilience, privacy, and access equity in contemporary library environments, this study investigates the role of Virtual Private Networks (VPNs) in fostering sustainable cybersecurity within the framework of Society 5.0 libraries. It does this by looking at the latest developments, applications, difficulties, moral dilemmas, and tactical methods associated with VPN deployment. Using peer-reviewed journal articles, conference proceedings, white papers, and policy documents published between 2010 and 2024, a methodical approach to literature review was used. The literature that bridges the fields of cybersecurity, library science, and Society 5.0 concepts was the main focus of the review. …