C-Day Program

Undergraduate and graduate projects: scale 0- 10 with 0 representing "Poor" and 10 representing "Exceeds Expectations"

• Successfully completed stated project goals and reported deliverables (0-10)
• Methodology/Approach: All required elements are clearly visible, organized, and articulated (0-10)
• Effective verbal presentation (0-10)
• Evidence of Rigor (0-10)
• Merit and Broader impact (0-10)

Games: scale 0 - 10 with 0 representing "Poor" and 10 representing "Awesome"

• TECHNICAL: Technically sound with appropriate visual & audio fidelity(0-10)
• GAMEPLAY: Engaging & Fun, with an intuitive UI. Rules of play are clear. Includes a win/lose state(0-10)
• ORIGINALITY: Sound, Art, Design, or Code(0-10)
• Evidence of Rigor (0-10)
• Merit and Broader impact (0-10)

* Project will be featured during the Flash Session

• UC-0180 Encoding Creative Commons Licenses to Images (Undergraduate Project) by Pellegrini, Tyler, Wilcox, Trey, Johnson, Marcus, Oberlin, Connor,
  Abstract: In our project we were tasked with modifying Gimp’s metadata editor to allow artists to check and add Creative Commons licenses and metadata to their image’s. This is done so that artist have an extra layer of protection for themselves and their art, with the ability to choose from multiple types of licenses allowing them to tailor this protection to the needs and desires they have for their artwork.
  Department: Software Engineering and Game Development
  Supervisor: Prof. Nick Murphy
  Poster

• * UC-0205 Enhancing GIMP’s User Experience: Addressing Community UI/UX Issues (Undergraduate Project) by Stanley, Alexander, Harrison, Ryan, Galvan, Dante, Elmostafa, Rami
  Abstract: Water quality monitoring is crucial for environmental protection, public health, and ecosystem sustainability. With increasing pressures from urbanization, agricultural runoff, and climate change, robust data-driven approaches are essential for early detection of water quality degradation and informed decision-making in environmental conservation efforts. Current water quality monitoring relies on reactive threshold exceedances, failing to detect gradual degradation and multi-parameter deterioration patterns. This creates delayed response to pollution events and missed opportunities for preventive intervention in one of Queensland's most vital water systems. The importance objective is to implement and evaluate a Real-Time Multi-Stream Monitoring system for early detection of water quality deterioration by monitoring key parameters simultaneously, using the Brisbane River data for baseline establishment and testing detection performance across multiple control limit configurations.
  Department: Software Engineering and Game Development
  Supervisor: Prof. Nick Murphy
  Presentation | Poster

• UC-0223 Predicting NBA Player Re-Injury Using Net Rating (Undergraduate Project) by Tention, Anaya,
  Abstract: This project examines whether player performance data can signal injury risk before an absence occurs. Using game-by-game net rating trends, I applied an exponentially weighted control-chart approach to detect early shifts in performance that might indicate a rising risk of re-injury. The method successfully identified 71% of re-injury cases with an average 20-game lead, suggesting that performance declines can serve as an early warning signal. While the false-alarm rate was high, the results show that performance-based monitoring has potential value for teams seeking proactive player-health insights.
  Department: Data Science and Analytics
  Supervisor: Prof. Michael Frankel, Dr. Austin Brown – Faculty Advisor
  Poster

• * UC-0224 Loving Arms: Website Audit & Redesign (Undergraduate Project) by Harrison, Marcus, Stropoli, Chris, Gibson, Katherine, Berger, Vaughn, Flores Valdez, Jesus,
  Abstract: Our team’s project aims to enhance the Loving Arms Cancer Outreach website, making it more accessible and user-friendly for all visitors. Loving Arms is a nonprofit supporting individuals and families affected by cancer, and their website plays a key role in sharing information, connecting people to support programs, and reaching those in need. To achieve this, we first audited the current site to identify areas of confusion or outdated content, and we are currently in the process of providing a redesigned website. These updates are intended to improve the visitor experience while giving staff a reliable, easy-to-use platform that better supports the organization’s mission.
  Department: Software Engineering and Game Development
  Supervisor: Dr. Yan Huang, Project Sponsor: Catherine Gankofskie - Loving Arms
  Presentation | Poster

• UC-0250 Tortured Artist (Undergraduate Project) by Tigani, Caitlin, Scholl, Ben, Tucker, Anaiya, Tucker, Adam,
  Abstract: You are a photographer that wants to move, so you take pictures of your house to give to your real- estate agent. However, as you are developing the photos you hear a noise that makes you turn on the lights, ruining your photos. Now you must retake the photos before morning, but something around the house has changed. Rooms are no longer in the right place, items are moved around, doors are locked, and an entity is watching you. Will you find the secrets within the puzzles or be left tortured?
  Department: Software Engineering and Game Development
  Supervisor: Dr. Joy Li
  Poster | More Information

• UC-0253 Stock Price Predictions Using LSTM & Technical Indicators (Undergraduate Project) by Elison, Kendal, Smith, Allen, Quinn, Dylan,
  Abstract: Stock price predictions using traditional statistical methods remains challenging due to market volatility and nonlinear dynamics. Long Short-Term Memory (LTSM) networks may model temporal dependencies in stock data more effectively than traditional statistical methods. Historical data for several companies’ stocks was obtained from Yahoo Finance, where it was then enriched with various technical indicators such as momentum and volatility. Preliminary analysis through Scala programming language suggests that incorporating these technical indicators can enhance short-term price prediction accuracy. Future works may seek to integrate additional trend and volume based indications in another, more robust, programming language like Python.
  Department: Computer Science
  Supervisor: Dr. Dan Lo
  Poster

• * UC-1140 RiverGuard - AI Powered Trash Detection (Undergraduate Project) by Versluis, Grant, Tucker, Collin, Bramblett, Wyatt, Pinto, Pedro, Ruiz, Geshlee,
  Abstract: RiverGuard’s mission is to protect and preserve waterways by using technology to identify and reduce pollution. The system uses an object detection model to automatically locate and classify trash within images or video of rivers and lakes, removing the need for slow, manual observation. By providing real-time insight into waste accumulation, RiverGuard helps communities, researchers, and organizations take faster, more effective action to keep waterways clean. Its goal is to create a sustainable monitoring system that empowers people to understand pollution patterns and support long-term environmental responsibility. RiverGuard represents a step toward cleaner water, healthier ecosystems, and a more informed approach to environmental protection.
  Department: Computer Science
  Supervisor: Prof. Sharon Perry
  Presentation | Poster

• UC-1149 AI Systems For a Stealth Videogame (Undergraduate Project) by Ehler, Lukas, Weir, Evan, Corona, Javier,
  Abstract: Shmovement Games 3/5 has been making progress towards creating a videogame that leverages multiple types of AI systems working in unison with one-another. This will be a game in which the player sneaks around enemies while trying to grab objects marked as the player's primary objective. This is a project for our CGDD 4242 AI class in which we are implementing a single agent, multi-agent, and advanced multi-agent system into one game.
  Department: Software Engineering and Game Development
  Supervisor: Prof. Nick Murphy
  Poster

• * UC-1152 SustainSync (Undergraduate Project) by El-Shaer, Youssef, Khan, Zaid,
  Abstract: Sustain Sync investigates how organizations can standardize sustainability tracking and how AI can convert that data into actionable insights. The platform normalizes utility data into a consistent schema aligned with industry frameworks. It applies machine learning forecasting and a retrieval-augmented co-benefit engine to relate sustainability goals across domains such as CO2, water, and biodiversity. Through a simple dashboard, Sustain Sync demonstrates an end-to-end data-driven approach for goal tracking, trend analysis, and AI-driven sustainability recommendations for an organization.
  Department: Computer Science
  Supervisor: Prof. Sharon Perry
  Presentation | Poster | More Information

• * UC-1166 Southern Bathtub Race: A Video Game Revival (Undergraduate Project) by Roberson, Caroline, Crouch, Jes, Gardner, Treonna, Landaverde, Jose, Lashley, Jake,
  Abstract: Southern Bathtub Race is a video game commemorating the annual SPSU Bathtub Races held from 1968 - 1991. The player controls a bathtub racer from a 1st-person perspective as they compete against the computer to navigate a racetrack modeled after the SPSU campus of 1991. The environmental assets and textures were created with traditional acrylic painting. The game is intended to replicate the races and student spirit on the SPSU campus for both SPSU alumni and KSU students.
  Department: Software Engineering and Game Development
  Supervisor: Dr. Yan Huang, Sponsor/Client: Will Mckenna
  Presentation | Poster | More Information

• UC-1168 Shepherd's Sin - A Visual Novel Hybrid Game Made with Unity (Undergraduate Project) by Randolph, Ara, Egl, Rin, Joiner, Everett, Swerdlow, Jonah,
  Abstract: By day, the grand old house shifts and shudders as if though alive. The six other residents gather in its lounges and parlors, sipping tea, squabbling over rooms, and faking civility. They laugh, they bicker, and they carry on as though nothing festers within these walls. When night falls, their facades rot away. They twist into monstrous embodiments of malice, each one a reflection of the seven deadly sins. By morning, they forget. You do not. Armed with a worn-out Monster Hunter’s Guidebook, you must reclaim its missing pages to learn who these people truly are, what they truly are. You must uncover their weaknesses and gather the weapons needed to destroy them before the storm clears.
  Department: Software Engineering and Game Development
  Supervisor: Prof. Nick Murphy
  Poster

• UC-1175 Admissions Assistant - AI Chatbot (Undergraduate Project) by Oduro, Chelsea, Bond, Jeffery, Morgan, Tanner, Scott, Allen, Parks, Joy,
  Abstract: The Kennesaw State University Graduate Admissions website contains extensive information on programs, application processes, deadlines, and eligibility requirements. However, its dense structure can make it difficult for prospective students to quickly locate specific details. This often leads to repeated inquiries from admissions staff, increasing their workload and delaying responses to more complex applicant needs. This project addresses those challenges by introducing a chatbot capable of retrieving and presenting the website information in a more intuitive, conversational format, improving user satisfaction and reducing workflow overhead for admissions staff.
  Department: Information Technology
  Supervisor: Prof. Donald Privitera, Christopher Kibbe - Project Sponsor
  Poster | More Information

• * UC-1183 MorphyxCam: Instant Photo Transformation Tool (Undergraduate Project) by Awatey, Priscilla, Doan, Long, Weng, Shaokun, Merchant, Aryan,
  Abstract: MorphyxCam is an interactive browser-based application that lets users capture live images and apply real-time visual effects. The system performs color filtering, shading adjustments, dynamic warping, and expressive distortion effects. Users can instantly reshape features, apply artistic styles, and wrap their photos onto 3D surfaces, creating engaging and playful visual transformations. By capturing live camera images and transforming them through pixel-level filtering, distortion effects, and 3D surface mapping, the system shows how multimedia techniques can be applied creatively within a web browser. Overall, MorphyxCam showcases the potential of interactive digital imaging and highlights how accessible web technologies can be used to build creative camera-based applications similar to modern photo apps.
  Department: Computer Science
  Supervisor: Dr. Alan Shaw
  Poster | More Information

• * UC-1196 Verocity: A Reactive Combat Framework (Undergraduate Project) by Moore, Brendan,
  Abstract: Verocity is a Minecraft plugin designed for fast-paced, visceral combat, where interactivity and FUN take center stage. The complex mathematics and system design required to build this plugin push the limits of standard Minecraft development, providing a reactive framework for advanced combat interactions. New Actions and Combat Features: Enhanced Basic Attacks – Smooth, responsive, and satisfying to chain together. Throwable Items – Every item can be thrown. Swords lodge into enemies on impact, ready to be recovered. Dashing – Lunge to swords stuck in the ground or at enemies to pull them out while tactically repositioning. Umbral Blade – Command a powerful blade that hovers behind you, responding to your inputs. Combat revolves around mastering its use. Grabbing Enemies – Pick up and throw foes, opening up new combos and utility in combat.
  Department: Computer Science
  Supervisor: Christopher Regan
  Presentation | Poster | More Information

• * UC-1207 AI Driven Resident Inquiry Processing (Undergraduate Project) by Moran, Ben, Sachwani, Sahil, Ashe, Thomas, Johnson, Sean,
  Abstract: The AI Driven Resident Inquiry Processing System is designed to enhance the National Housing Compliance (NHC) ability to process resident inquiries using artificial intelligence(AI). NHC is a 501(c)(4) not-for-profit corporation who provides training and compliance services to the affordable housing industry. Each month NHC receives over 200 inquiries from residents via phone and email. These inquiries range from general questions to urgent, life-threatening concerns. Efficiently processing and responding to these inquiries is often critical to resident safety and well being. This project uses AI to automate resident inquiries as they are received, extract and classify key information, and display this information on an interactive dashboard for close monitoring and timely resolution. By streamlining inquiry processing, this system enables NHC to respond to resident needs faster and accurately.
  Department: Information Technology
  Supervisor: Prof. Donald Privitera, Jeff Wirrick- Project Sponsor, Kelli Sterling- Project Sponsor: Walter Hoang- NHC IT staff, Jamaul Morrison- NHC IT staff, Xeryus Starr- NHC IT staff, Satara Tyler- NHC staff
  Presentation | Poster

• UC-1211 Machine Learning Linux Log Anomaly Detection (Undergraduate Project) by Scott, Samuel, Loisy, Audrey, Silva-Rivas, Dylan, Brady, Sheamus,
  Abstract: Cybersecurity is becoming an increasingly important part of digital life. Malware can silently intrude on a user’s system and perform malicious actions and generate unusual system behavior without the user ever being aware. This malware often presents with unusual system logs being generated. These logs, however, are difficult to consistently track and analyze, especially for casual users. To help bridge this gap between hard-to-read log data and the useful information it contains, we created LUAADS (short for Linux User Account Anomaly Detection System), designed for Ubuntu systems. LUAADS can automatically collect entries from common log files (such as syslog and auth.log), parse them into an easier-to-read format, and then analyze them for system patterns using machine learning. LUAADS can automatically alert the user when a log entry is anomalous and offers a feedback mechanism to improve on any false positives. LUAADS also offers a user-friendly GUI that allows non-tech-savvy users to be able to find and sort all their system logs in a single location. By bringing analysis of system logs to a wider audience, LUAADS helps improve Linux system security, even for non-tech-savvy users.
  Department: Computer Science
  Supervisor: Prof. Sharon Perry
  Poster

• UC-1222 Active Learning System for Labeling Chest X-rays (Undergraduate Project) by Hall, Matthew, Lane, Noah, Smith, Josh, Merrill, Elijah,
  Abstract: This project aims to develop a complete Active Learning System for chest X-ray image classification, designed to automate data preparation, streamline model training, and reduce the manual effort required for medical image labeling. The system establishes a structured and scalable pipeline that moves from raw data ingestion to automated decision-making, incorporating dataset indexing, patient-aware splitting, preprocessing, configuration management, and validation to ensure data flows reliably through the system. The model component uses CNNs to generate baseline diagnostic predictions across chest pathologies. Active learning strategies are then applied to identify the most informative unlabeled images, enabling iterative retraining that improves model performance while minimizing labeling cost. Proposed strategies for training improvements include transfer learning with a pretrained model. Evaluation tools such as a dashboard to track performance would also enable a reproducible, clinically relevant image workflow. The final system would deliver a reproducible framework capable of managing large medical imaging datasets, selecting high-value samples for annotation, and continuously refining classification accuracy over time.
  Department: Computer Science
  Supervisor: Prof. Sharon Perry
  Poster | More Information

• UC-1226 iKnowIT: Multilingual Smartphone Tutorial Platform (Undergraduate Project) by Juarez, Jacqueline, Bazan, David, Rivera, Julissa,
  Abstract: Digital literacy challenges affect millions of adults who struggle with basic smartphone use due to rapidly changing technology and limited support. iKnowIT is a dynamic, web-based learning platform designed to provide clear, visual, and multilingual tutorials that guide users through essential device functions. The goal of iKnowIT is to bridge the digital divide and empower users to engage confidently with modern technology
  Department: Computer Science
  Supervisor: Prof. Sharon Perry
  Poster

• UC-1232 Stronghold (Undergraduate Project) by Tigani, Caitlin, Tucker, Adam, Lloyd, Camden, Balsor, Dale,
  Abstract: Our goal for this game is to create a game using both a single agent and multi agent AIs. In our game you can either play against the current AI or train the AI up as it fights against another AI. The training will use genetic AI. So whichever AI wins that row will be the one that move on. The loser will have their weights adjusted. This means the more that you train the Stronghold AI the harder the AI will be to fight against. The final game mode available is being able to play against another.
  Department: Software Engineering and Game Development
  Supervisor: Prof. Nick Murphy
  Poster

• UC-1234 Fundoria (Undergraduate Project) by Stabler, Connor, Jones, Dameon, Hassan, Rafay, Liu, Michael, Horne, Brenden,
  Abstract: Fundoria is an educational technology project designed to transform financial literacy and career readiness for elementary and middle-grade students. The platform combines interactive gameplay, AI-driven personalized learning, and real-world job simulations to teach essential financial concepts in an engaging, age-appropriate way. Students explore budgeting, saving, banking, and career skills through guided missions that adapt to their learning styles. Fundoria also provides teachers and schools with standards-aligned activities, classroom resources, and assessment tools that connect academic content to practical, everyday decisions. By blending gamification with evidence-based instruction, Fundoria empowers students to build confidence, develop responsible financial habits, and prepare for future academic and career success.
  Department: Software Engineering and Game Development
  Supervisor: Prof. Yan Huang
  Poster

• * UC-1240 Project Ibis (Undergraduate Project) by Alderman, Isaac, Mizell, Braden, Sutton, Collin,
  Abstract: Project Ibis (working title) is a first person, narrative heavy, puzzle-lite RPG that follows the story of a renaissance era plague doctor and their attempt to alter the minds of three subjects; a gardener, a street urchin, and a priest. The game’s narrative is set in historically accurate 1637 Florence, Italy, in the wake of the Great Plague of Milan, and draws heavily from renaissance culture. Each subject being treated is a complex person with personal conflicts and issues; our focus is on tackling mental and emotional health through empathy and nuance rather than diagnosis. Different aspects of each subject’s psyche take physical form in the dreamscape environments, allowing the player to speak and interact with them directly, influencing the state of the subject themself, and, in turn, the environment around the player. The game features many puzzles, with solutions ranging from alchemy, exploration, and dialogue, to spatial and logical reasoning. The primary goal of play is to reach the core of each subject’s mind, allowing the opportunity to change their fundamental outlook on life, if the player so chooses. Either way, the consequences will follow the player until the end. Our primary goal with this game is to make our players think carefully about how their actions might affect each patient they treat, and find their own answer to the question of how much they should influence each one.
  Department: Software Engineering and Game Development
  Supervisor: Dr. Joy Li
  Poster

• * UC-1244 Agentic AI For Intelligent Customer Communication (Undergraduate Project) by Papadopoulos, Lucas, Hopkins, Jeremy, Gargour, Munir, Dease, Weston,
  Abstract: E-commerce web shoppers need fast, reliable responses to a variety of requests: account modifications, order tracking, or policy inquiries. Businesses must address user queries in a fast and efficient manner, or else lose customers. Multi-agent AI models boast the ability to answer customer questions and act upon consumer queries without outside intervention. However, research is sparse as to how agentic models can transfer benefit to large commercial software stacks under realistic commercial load. We sought to ask whether a multi-agent AI architecture can effectively handle commercial-scale e-commerce customer service tasks. Moreover, we investigated how a multi-agent AI architecture compares to traditional single-agent customer service solutions in handling complex e-commerce tasks. Our team developed a multi-agent AI architecture using specialized Claude Haiku agents coordinated through LangGraph, with a React frontend and PostgreSQL DB-Kafka backend. Testing will compare our multi-agent system against a single-agent baseline to evaluate effectiveness in handling complex customer service requests. Preliminary results have shown that an agentic AI architecture significantly increases query-response correctness for customer requests.
  Department: Computer Science
  Supervisor: Prof. Sharon Perry, Project Sponsor: Capgemini - Anna Richard and Senthilraj Duraisamy
  Poster | More Information

• * UC-1257 Integrating a Pressure Profiling Dock into an Open-Source Digital Art Framework (Undergraduate Project) by Miller, Luke, Enyart, Noah, Smith, Jonah, Jones, Jaeden,
  Abstract: The effectiveness and expressiveness of digital drawing are heavily dependent on an artist’s control over their medium; pen pressure dynamics, which enables fine-tuned variation in line weight and opacity. While the use of this tool improves an artist's control, the open-source editor GIMP currently offers only limited means to visualize or adjust these pressure behaviors in real time. GIMP does support pressure-based input, however, it lacks a dedicated interface for configuring a personalized pressure profile, creating a barrier for users seeking precision comparable to commercial tools. This project addresses that gap by designing and implementing a Stylus Pressure Profiling Dock within the GIMP application environment. in addition to the implementation, this project also investigates key research questions, including how pressure data can be effectively captured from various drawing tablet devices within GIMP’s existing input architecture, and how optimization techniques can be combined with the logistic (sigmoid) function to generate an accurate, user-customizable pressure S-curve.
  Department: Software Engineering and Game Development
  Supervisor: Prof. Nick Murphy
  Presentation | Poster

• UC-1259 Light'em Up (Undergraduate Project) by Sutton, Collin, Jones, Ronnie, Anderson, Max,
  Abstract: The primary goal of “Light’em Up” is to create engaging and intelligent AI that can operate within three degrees of freedom and against forces of gravity. Enemies will track the player, predict their movement, and collaborate to set traps and outflank them. All of this takes place in space, at high speeds, and at a scale where gravity has a real effect on navigation. We have four distinct AI enemies at play: Homing missiles - single agent system that follows the player’s movement at a slightly faster speed Tracking missiles - single agent system that moves at a constant speed and direction based on prediction of the players position at the time of impact/detonation Local Space Force - Multi-agent system that follow the player while firing lasers and avoiding collision with obstacles Inter-Galactic Space Force - Advanced multiagent system that corporate in order to put the player in disadvantaged positions while also inheriting behaviors from the Local Space Force
  Department: Software Engineering and Game Development
  Supervisor: Prof. Nick Murphy
  Poster | More Information

• UC-1260 Digital Micromouse Maze Simulation (Undergraduate Project) by Bell, Chase,
  Abstract: We construct a digital simulation of the Micromouse competition and analyze popular algorithms for searching (A*, Dijkstra’s, and Flood-fill). From our analyses, we design an algorithm with the goal of achieving the shortest possible run time to the end of the maze.
  Department: Software Engineering and Game Development
  Supervisor: Prof. Nick Murphy
  Poster

• UC-1261 AI-Powered GRE Vocabulary App (Undergraduate Project) by Verde, Michael, Landeta, Ellyan, Onuorah, Cynthia, Tran, David, Binchamo, Bereket,
  Abstract: Our project develops a client-side React application for GRE vocabulary practice using structured JSON word data. The site supports filtering, search, audio output, and randomized quizzes. A reinforcement-learning hint system, inspired by prior research on adaptive learning, guides users toward difficult vocabulary. We aimed to create an interface that demonstrates how lightweight front-end tools can support personalized study without requiring a backend.
  Department: Information Technology
  Supervisor: Prof. Donald Privatera, Manohair Sai Jasti (Project Sponsor Founder of Scafwording)
  Poster | More Information

• UC-1262 Georgia Watch - Georgia Hospital Accountability Score (Undergraduate Project) by Straiton, Robert, Cox, Patrick, Nortey, Constant, Amaravadi, Sankalp, Keller, Kahmin,
  Abstract: Georgia Watch and the members of this team have partnered to change how Georgia residents understand their healthcare by providing a source of objective metrics which affect their care. This project represents the quintessential React Project produced with the purpose of a reactive interface for the end user, built for maintainability for any subsequent developers. The hospital data is maintained in JSON format for its ease of parsing and adjustment pending any changes. The interactive map was developed through the MAPBOX library, and the team is maintaining a deployment via an independent repository with necessary control over the website.
  Department: Software Engineering and Game Development
  Supervisor: Dr. Yan Huang
  Poster | More Information

• UC-1263 BudgetWise - The College Friendly Budgeting App (Undergraduate Project) by Thompson, Taylor, Issa, Yasmeen, Khan, Sameer, Nguyen, John, Lechuga, Reynaldo,
  Abstract: For our senior project, we developed BudgetWise, a budgeting app designed to help college students manage their finances with confidence. BudgetWise has an emphasis on ease of use and accessibility, with features such as dark mode for improved visibility. Bank accounts and credit cards can be securely linked to the user’s account where they can track their recent purchases, create budgets based on their personalized needs, and track their spending with a dynamic progress bar that changes colors the closer they get to their budget limit. By combining financial tools with accessibility, BudgetWise empowers students to make informed financial decisions and help build strong money management habits.
  Department: Computer Science
  Supervisor: Prof. Sharon Perry
  Presentation | Poster

• UC-1268 Falling Debris (Undergraduate Project) by Duarte, Julian, Martin, Adam, Collins, Rylan, Haggard, Hugh, Boecker, Chance,
  Abstract: Falling Debris is a student-developed party 2D platformer game in which players have to survive falling blocks by grappling upward to outlive the others. The game is played in rounds, and between them, players can purchase items to improve their odds of survival.
  Department: Software Engineering and Game Development
  Supervisor: Dr. Joy Li
  Poster

• UC-1269 Website Makeover - Loving Arms Cancer Outreach (Undergraduate Project) by Pacheco, Yuliana, Cook, Alicia, Tompkins, Taylor, Bridges, Schuyler, Roberts, Dalton,
  Abstract: Loving Arms Cancer Outreach (LACO) provides financial, emotional, and community support to individuals affected by cancer, making an accessible and reliable website essential to its mission. Our team conducted a quality assurance audit using tools such as Google Lighthouse and axe DevTools, identifying issues with accessibility, navigation, readability, and mobile responsiveness. Using these findings, we redesigned key sections of the site, improved layouts and forms, and recommended updated plugins to enhance usability and long-term performance. We also developed a Website Architecture and Maintenance Guide to support sustainability. This project establishes the foundation for a modern, user-friendly website that strengthens LACO’s outreach and donor engagement.
  Department: Information Technology
  Supervisor: Prof. Donald Privitera, Cat Gankofskie (Project Sponsor)
  Poster | More Information

• UC-1273 V.A.P.R. Rush (Undergraduate Project) by Collins, Rylan, Duarte, Jullian, Hugh, Oliver, McMillian, Ethan,
  Abstract: A 3D platformer where you can transform from a cube to a boat and a plane. The game is on mobile and features the player traversing through a vapor wave inspired level with techno music in the background. They must perform jumps and lane switches to the beat of the song, and survive to the end of the level to win.
  Department: Software Engineering and Game Development
  Supervisor: Dr. Sungchul Jung
  Poster

• UC-1274 Cloud-Native CI/CD Pipeline (Undergraduate Project) by Meduteni, Enitan, Elmostafa, Rami, Crowley, Matt, Fleming, Kade, Graffree, Cecily,
  Abstract: This project documents a 12-week capstone implementing a cloud-native CI/CD pipeline using industry-standard DevOps tools. The system integrates Jenkins for continuous integration, Kubernetes for container orchestration, GitOps (ArgoCD) for automated deployments, DevSecOps practices including RBAC and vulnerability scanning, and comprehensive monitoring using Prometheus and Grafana. Mentored by Sudheer Amgothu, Principal Cloud Operations Engineer.
  Department: Software Engineering and Game Development
  Supervisor: Dr. Yan Huang, Industry Sponsor: Sudheer Amgothu
  Poster | More Information

• UC-1276 CI-CD Pipeline Team 2 (Undergraduate Project) by Arnold, Cameron, , , , ,
  Abstract: Our project is about creating a basic cloud-native pipeline that can build and deploy an application in a more automated way. We will also try to add some security checks and monitoring tools so that we can see how everything is working. The goal is to get hands-on experience with the process and show a working demo at the end of the semester.
  Department: Information Technology
  Supervisor: Prof. Donald Privitera, Mentor - Sudheer Amgothu
  Poster

* Project will be featured during the Flash Session

• GC-0151 Smart HR Onboarding with Microsoft 365 Team 2 Capstone Project (Graduate Project) by Gregory, Annalise, Colley, Michael, Laurent, David, Agarwal, Harshita, Kumar, Nilesh,
  Abstract: The Smart HR Onboarding with Microsoft 365 Capstone Project aims to transform the new hire onboarding experience by leveraging Microsoft 365 tools to deliver a seamless, automated process. Using Power Automate, we streamline tasks through automated emails, reminders, and documents sharing. This ensures every step of the onboarding journey is efficient and consistent. Sharepoint and Outlook serve as the main sources of storing and sharing required tasks. Power Automate's integration with Power BI allows for real-time reporting with an interactive dashboard the provides insights into onboarding progress or areas where new hires need support. This project aims to simplify onboarding for HR teams and makes onboarding progress measurable.
  Department: Information Technology
  Supervisor: Dr. Jack Zheng
  Presentation | Poster | More Information

• GC-0157 AI Graduate Admissions Assistant (Graduate Project) by Hyatt, Katherine, Wright, Ginger, Edgar, Pablo, Akinwusi, Oluwatosin, Johnson, Alan,
  Abstract: Project Overview: Create a "Proof of Concept" for an AI Graduate Admissions Assistant chatbot that will: • Autonomously reference KSU website information in real-time • Direct users efficiently to specific, relevant content • Reduce questions for admissions personnel • Self-update its knowledge base when website content changes •Provide accurate, instant responses to common admissions questions and allow for response correction in an admin dashboard
  Department: Software Engineering and Game Development
  Supervisor: Dr. Reza Parizi
  Poster

• GC-0241 Using Dog Breed Classification Uncertainty Estimation to Inform Mixed Breed Ancestry (Graduate Project) by Mackey, Brandon, Koya, Maryam, King, Theodore, Hutchison, Scott,
  Abstract: This study investigates whether Monte Carlo uncertainty estimation and probability distributions can be used to identify ancestral composition of mixed-breed dogs. A dataset containing images of purebred dogs was used to train a Monte Carlo Dropout model. The trained model will next be tested on images of mixed breed dogs. Our hypothesis is that the model can be used to provide informative probability distribution for breed ancestry classification, offering a potentially valuable tool for analyzing the genetics of dogs.
  Department: Computer Science
  Supervisor: Dr. Coskun Cetinkaya
  Poster

• GC-0251 Reproducing Extended Isolation Forests with STAR-CAST (Graduate Project) by Patrick, Drew, Sivakoti, Ram Sai, Malik, Rohit, Kamal, Vardhineedi Surya, Palagundla, Venkata Sasidhar Reddy,
  Abstract: Fraud models routinely flag suspicious transactions but rarely explain why, which slows investigations and erodes trust. In this work we study Extended Isolation Forest (EIF) for unsupervised fraud detection and propose STAR-CAST, a lightweight framework that turns raw anomaly scores into threshold-aligned IF–THEN rule cards with explicit reliability measures. Using the public credit-card fraud dataset (284,807 transactions, 492 frauds; ~0.17% prevalence), we apply a time-aware 70/15/15 Train/Validation/Test split and fit-on-train preprocessing (Amount log1p→z; Time z; V1–V28 retained). We train IF, EIF, an EIF ensemble, a Mahalanobis baseline, and density models (HBOS, COPOD, ECOD) fully unsupervised, evaluate them as rankers first (PR-AUC, Max-F1, Precision@K), then freeze Validation thresholds and measure Test-set performance at FPR-targeted and recall-targeted operating points, with hourly block-bootstrap CIs to capture uncertainty. On Test, density/tail methods perform best on these PCA-style features (HBOS AP ≈ 0.139, COPOD ≈ 0.103, ECOD ≈ 0.072), while EIF consistently improves over IF (EIF AP ≈ 0.0534, EIF-ENS ≈ 0.0537, IF ≈ 0.0498; Mahalanobis ≈ 0.0456). At a realistic FPR of 0.5%, HBOS and COPOD achieve usable precision and recall, while EIF remains operationally competitive. STAR-CAST generates compact rules with high stability, fidelity to the model decision, and validated local precision, providing auditable, actionable explanations rather than opaque scores.
  Department: Computer Science
  Supervisor: Dr. Dan Lo
  Poster

• * GC-0258 SafeCircle: AI and Micro-radar-based Remote Monitoring for Patients with AD/ADRD (Graduate Project) by RAHMAN, AWAN-UR-, Borty, Soarov, Quddus, Shakib,
  Abstract: Alzheimer's disease and related dementias (AD/ADRD) is an irreversible and degenerative neurological condition that severely impacts neurons, resulting in cognitive decline and memory loss. This study explores a mHealth system, including a SafeCircle iOS prototype, a novel solution that combines artificial intelligence with cutting-edge micro-radar technology. The platform offers a variety of features, including management of patient and caregiver profiles, real-time alerts in case of emergencies, emergency contact lists, one-touch SOS support, sharing of live locations, and recording of unusual events in video. It is a responsive and reliable care assistant that optimizes patient safety while reducing caregiver burden.
  Department: Information Technology
  Supervisor: Dr. Nazmus Sakib, Dr. Sumit Chakravarty
  Presentation | Poster | More Information

• * GC-0270 OncoBoost - Hydration Monitoring Application (Graduate Project) by Madubike, Blossom, Alam, Aafra, Ojo, Damola,
  Abstract: Dehydration is a common and preventable complication for oncology patients, especially those undergoing chemotherapy and radiation. Side effects such as nausea, fatigue, and loss of appetite make it difficult for patients to maintain adequate fluid intake, contributing to avoidable discomfort and potential treatment disruptions. This capstone project presents Onco-Boost, a mobile hydration monitoring application designed to help adult oncology patients track daily fluid intake, recognize their intake patterns, and stay engaged in daily self-care between clinic visits. Built with React Native and Expo, and backed by Firebase for authentication and cloud data storage. Onco-Boost translates clinical hydration guidance and research into an intuitive, patient-friendly interface. The app features a streamlined home page for quick-tap beverage logging with 4 beverage options, a history page that displays previous entries and daily totals, and a profile page where users can set or adjust basic information. The system is designed with simple data models, role-appropriate access, and secure storage to support future integration with clinical workflows if desired. Onco-Boost demonstrates how a focused, data-driven mobile tool can support oncology patients in maintaining hydration and may reduce preventable complications related to low fluid intake.
  Department: Information Technology
  Supervisor: Dr. Zhigang Li, Dr. Tracy Ruegg, Dr. Jack Zheng
  Presentation | Poster | More Information

• GC-1144 Meetless: The Operating System for Business in the AI Age (Graduate Project) by Pham, An,
  Abstract: Meetless functions as an OS for modern organizations: a secure, multi-agent runtime that turns scattered inputs (docs, emails, tickets, chats) into asynchronous, outcome-driven discussions with decisions, owners, and due dates—so teams ship without meetings. System architecture (OS metaphor).- Kernel (Orchestrator): schedules “processes” across specialized agents using routing policies, guardrails, and retry semantics. - Process I/O: unified connectors for Google/Microsoft suites, Slack, Jira/Linear, and web sources, normalized into a document graph with vector embeddings. - Memory & FS: temporal knowledge graph (Neo4j) + document store (PostgreSQL) + vector index (Weaviate) with time-aware retrieval. - Syscalls/APIs: /discussions.create, /decisions.propose, /actions.sync, /summaries.latest; all idempotent with audit trails. - Runtime & Toolchain: FastAPI services; LangGraph/LangChain for agent flows; MLFlow for experiment tracking; background workers for ingestion, scoring, and SLA enforcement. Workflow. 1. Ingest context: generate a structured Brief (goals, constraints, risks, dependencies).2. Facilitate threaded Q&A; auto-summarize positions and trade-offs. 3. Produce a Decision & Action Pack (decisions, owners, due dates) and sync to PM tools. 4. Update the temporal graph; surface follow-ups and unresolved questions. Security & governance: RBAC with tenant isolation, secret-scoped connectors, redaction/masking policies, and full decision lineage via timeline views. Evaluation targets (pilot teams, n=6): ≥40% reduction in meeting hours for standups/planning/retros; <5 min median catch-up time for late joiners; measurable increase in decision traceability and follow-up closure rate.
  Department: Computer Science
  Supervisor: Dennis Loubiere
  Poster | More Information

• GC-1146 Student Engagement Portal: Enhancing Student Success through Milestone Tracking (Graduate Project) by Brewer, Antonio, Bolinger, Taylor, Dawkins, Tyler, Franck, Ricartho, Valles, Moises,
  Abstract: The Student Engagement Portal, also known as the Milestone Map, is a platform developed to help students within KSU’s College of Computing and Software Engineering monitor their academic and professional growth. The system enables students to log milestones, check in at events, and view progress toward personal and departmental goals. Built with a full-stack architecture using NestJS, React, and MongoDB, the portal also includes an administrative dashboard for event management and analytics. The project demonstrates how progress tracking and clear visualization of achievements can improve communication, organization, and engagement between students and the college.
  Department: Software Engineering and Game Development
  Supervisor: Dr. Reza Parizi
  Poster

• * GC-1147 Enhancing Mass Casualty Triage Training Through Human–AI Collaboration in Virtual Reality (VR) (Graduate Project) by Aiyatham Prabakar, Rishi Kiran,
  Abstract: Mass casualty triage requires quick, accurate decisions under pressure. Live training is costly and time-intensive. Virtual Reality (VR) trainings approach has shown comparable learning effectiveness compared to live trainings, motivating the use of VR simulations. This study explores how collaboration with an AI robot partner can enhance the triage training effectiveness. The Findings will contribute in understanding how human-AI collaboration enhances trainings.
  Department: Information Technology
  Supervisor: Dr. Hansol Rheem
  Presentation | Poster

• * GC-1154 Peer Evaluation Automation and Feedback System (Graduate Project) by khan, sameer, Okafor, Nnedi,
  Abstract: A web-based platform to streamline peer evaluations in team-based courses. Professors can securely create and manage student rosters, assign students to courses/teams, trigger email invitations, and receive structured, professor- friendly reports with both numeric and textual feedback. Optional AI features may summarize comments and flag potential concerns, depending on timeline and scope.
  Department: Information Technology
  Supervisor: Dr. Jack Zheng, Project Advisor: Geetika Vyas
  Presentation | Poster | More Information

• GC-1162 Present Panic Game (Graduate Project) by Adegeye, Adetunji, Bryant, Kendrick, Ayeronwi, Catherine,
  Abstract: Present Panic is a festive top-down 2D arcade game where players control an elf inside Santa’s workshop and must collect presents while avoiding Christmas-themed enemies. During Sprint 2 in our SWE class, our team transitioned from a functional prototype to a fully developed Beta, expanding assets, implementing UI systems, building levels, and integrating polished game mechanics following the MDA framework. We developed new artwork, sound effects, animations, and user interface elements including a main menu, HUD, pause menu, and scene transitions. A survey and both manual and automated testing were prepared to gather user feedback. This Beta version forms the foundation for the final game release, to be improved based on testing and player experience.
  Department: Software Engineering and Game Development
  Supervisor: Prof. Brooke Zhao
  Presentation | Poster | More Information

• * GC-1198 Onboarding Tool for New Smartphone Users (Graduate Project) by Velagapudi, Namita, Dasari, Bhagya Surekha, Maddineni, Yaswanth, Chekka, Rohith Venkata Sai, Pinninti, Balachandar,
  Abstract: The Smartphone Onboarding Tool is an interactive web platform created to help seniors and new smartphone users become comfortable with mobile technology. It offers a realistic, simulated smartphone interface, guided walkthroughs, and an easy-to-use design that builds confidence in performing everyday tasks. Caregivers can monitor user progress, while learners can practice safely without affecting an actual device. The solution is developed with a React frontend, a Node.js/Express backend, and an SQLite database, all built with a strong focus on mobile-first accessibility.
  Department: Information Technology
  Supervisor: Dr. Jack Zheng, Mentor: Robert Thompson
  Poster | More Information

• * GC-1215 ClinicalRAG: A Scalable Benchmark of Privacy, Relevance, and Speed in Semantic Retrieval for clinical transcriptions (Graduate Project) by Kumar, Pradyumna, Dandibhatla, Sai Sruti, Subramanian, Srinivasan, Anukula, Purna Chandu, Athukuri, Pranitha,
  Abstract: Traditional keyword search struggles with the scale, complexity, and contextual depth of clinical data. This project develops and evaluates semantic search systems that better understand medical language, enabling physicians and researchers to retrieve contextually relevant information through a Retrieval Augmented Generation (RAG) framework. We integrate privacy-preserving methods, including differential privacy and homomorphic encryption to protect sensitive clinical transcriptions. For improved speed and accuracy, we enhance the baseline RAG architecture with Hierarchical Navigable Small World (HNSW) indexing and Maximal Marginal Relevance (MMR) based reranking. To ensure scalability, clinical documents are ingested using PySpark and stored in a vector database optimized for high-dimensional queries, enabling fast, accurate, and privacy-aware retrieval of medical transcriptions.
  Department: Computer Science
  Supervisor: Dr. Dan Lo
  Poster

• GC-1233 An AI-Powered Convolutional Neural Network System for Multi-Class Image Classification of Rice Plant Leaf Diseases (Graduate Project) by Buddharaju, Akshay Krishna Varma, Yellu, Siri, Peddi, Pranay Kumar,
  Abstract: This project focuses on building an intelligent system that can automatically identify common diseases found on rice leaves by analyzing simple images. Using a deep convolutional neural network, the model learns to recognize visual patterns associated with three major diseases: Bacterial Blight, Brown Spot, and Leaf Smut. These diseases often show subtle differences in color, texture, and leaf damage, and the model is trained to distinguish them accurately. The goal of this work is to show how artificial intelligence can support modern agriculture by helping farmers detect problems early, even without expert knowledge. By processing images through careful preprocessing, augmentation, and feature extraction, the system becomes capable of making reliable predictions from real-world leaf photographs. This reduces manual dependency, improves diagnostic speed, and promotes healthier crop management.
  Department: Computer Science
  Supervisor: Dr. Coskun Cetinkaya
  Poster

* Project will be featured during the Flash Session

• UR-0199 Blood Flow Simulation From Coronary Computed Tomography Angiography Using Vnet (Undergraduate Research) by Pinto, Pedro,
  Abstract: Coronary artery disease (CAD) is one of the leading causes of death worldwide, making the assessment of blood flow and pressure distribution within coronary vessels essential for diagnosis and treatment planning. Fractional Flow Reserve (FFR) is a key measure used to determine the severity of arterial blockages, but traditional methods, such as invasive measurements or computational fluid dynamics (CFD) simulations, are not only time-consuming and costly but also invasive. This project explores the use of deep learning to predict blood pressure distribution in coronary arteries using a 3D convolutional neural network. The dataset consists of Coronary Computed Tomography Angiography (CCTA) scans paired with blood pressure obtained from CFD simulations. After preprocessing and voxelizing the CCTA scans, the VNet model learns to map the vessel’s geometry to its internal pressure distribution. Our model achieved a Pearson correlation of 0.93 and an R² score of 0.84 between the predicted and simulated pressures. These results indicate that VNet delivers accurate, spatially consistent pressure predictions that closely match CFD outputs while substantially reducing computation time. This approach underscores the potential of deep learning to accelerate non-invasive FFR estimation and enhance patient-specific cardiovascular analysis.
  Department: Computer Science
  Supervisor: Dr. Chen Zhao
  Poster

• UR-0225 Carbonyl Detection in IR Using Deep Learning (Undergraduate Research) by Shakthivel, Dharani,
  Abstract: A carbonyl group consists of a C=O double bond, which produces a distinct spike in the infrared spectrum that can be recognized visually near 1700 cm-1. The dataset included spectra prepared using KBr disc, nujol mull, CCl4 solution, and liquid film methods. Liquid film and CCl4 solution produced the cleanest peaks because the analyte is measured without being diluted into a solid matrix, closely matching the way chemists typically identify carbonyls by eye. In contrast, the KBr disc and nujol mull techniques require mixing the sample with additional materials, which contribute their own absorbances and distort peak intensities, introducing interference that complicates automated detection.
  Department: Computer Science
  Supervisor: Prof. Maxwell Bradley
  Poster | More Information

• UR-0227 Compilation of Binary Neurons to OBDDs (Undergraduate Research) by Boyce, Aidan,
  Abstract: A neuron with binary inputs & outputs corresponds to a Boolean function. To explain and verify the behavior a neuron (and by extension, a neural network), we can explain and verify its Boolean function. There has been recent interest in representing the Boolean function of such a neuron as an Ordered Binary Decision Diagram (OBDD), which facilitates such analyses. We propose an algorithm for compiling a binary neuron into an OBDD using a compiler that decomposes a Boolean function into a decision graph. We augment this compiler so that it outputs an OBDD instead. Our augmented compiler produces intermediate OBDDs that represent inner- and outer-bounds of the original neuron, tightening compilation progresses. Theoretically, decision graphs of binary neurons are more succinct than their decision trees. Empirically, compilation to decision graphs can scale to neurons with over a thousand features, compared to dozens of features using other compilers.
  Department: Computer Science
  Supervisor: Dr. Arthur Choi
  Poster

• UR-0246 Quantum ML for Science & Engineering (Undergraduate Research) by Shakthivel, Dharani, Tan, Haoxian, Martin, Justin,
  Abstract: Classical machine learning methods such as CNNs, SVMs, PCA, Logistic Regression, and Random Forests, have achieved strong performance across fields from computer vision to drug discovery. However, these models face scalability limits when trained on large or high-dimensional datasets. Quantum computing introduces quantum mechanics like superposition, interference, and entanglement - enabling quantum kernels, quantum feature maps, and hybrid quantum-classical architectures that may reduce computational cost or enhance data representation. This project implements classical versions of these algorithms alongside their quantum counterparts to evaluate differences in accuracy and efficiency. By comparing performance across diverse scientific and engineering datasets, the study assesses whether quantum-enhanced methods can match or surpass classical baselines under practical constraints.
  Department: Computer Science
  Supervisor: Prof. Sharon Perry, KSU Supervisor: Dr. Yong Shi
  Poster | More Information

• UR-0247 Lyric Prediction Model (Undergraduate Research) by Gideon, Evan, Dasher, Garrett, Batanado, Ulrich, Claerbout, Drew, Henshaw, Andrew,
  Abstract: Word prediction plays a central role in the development and refinement of large language models, supporting applications such as search optimization, dialect identification, and conversational AI systems like Siri as AI text generation becomes increasingly widespread, the demand for precise and contextually aware predictive capabilities continues to grow. This project presents lyric prediction model designed to generate the next lyric based on preceding words, with the ability to identity line breaks and sequential structure. Ultimately, this work aims to advance lyrical text generation by enabling the model to emulate the stylistic characteristics of specific artists or musical genre.
  Department: Computer Science
  Supervisor: Dr. Dan Lo
  Poster

• UR-0248 Shell Commands Used in Cybersecurity Training (Undergraduate Research) by Kourk, Nathan, Suda, Jacob, Butler, Ming, Alemu, Yonnas,
  Abstract: This project explores patterns in shell command usage during cybersecurity training programs. We will analyze syntax frequency and selection of shell commands across multiple datasets looking for patterns in user behavior. The goal of this study is to identify differences between programs, and to provide insight into the trends within the command line environment.
  Department: Computer Science
  Supervisor: Dr. Dan Lo
  Poster

• UR-1163 Editing Films and Movies on GIMP (Undergraduate Research) by Al Saad, Ahmed, Grace, William, Robinson, Brian,
  Abstract: This project is to develop an import feature for Digital Picture Exchange (.DPX) files with GIMP. The DPX format is popularly used in the film and visual effects industry. DPX stores uncompressed high bit depth images. Although the project is still in progress, a working prototype is provided.
  Department: Software Engineering and Game Development
  Supervisor: Prof. Nick Murphy
  Poster

• UR-1167 OwlExchange (Undergraduate Research) by Chaplin, Laura, Johnson, Caitlin, D'Haiti, Alexin, Gordon, Jamaul, Barranco, Leonardo,
  Abstract: OwlExchange is a secure, role-based campus marketplace designed to help Kennesaw State University students buy, sell, exchange, and donate items in a trusted digital environment. The platform replaces unorganized and unsafe exchanges occurring across group chats and social media by providing a centralized system with authenticated user accounts, item listings, and transparent communication. Built with a Flask (Python) MVC architecture, Auth0 for secure identity management, and MySQL for persistent storage, OwlExchange supports modular dashboards for buyers, sellers, and administrators, enabling item management, interest requests, and platform oversight. By promoting reuse and donation of textbooks, furniture, electronics, and other student items, the system contributes to affordability and sustainability on campus. The project demonstrates full-stack development, secure authentication design, UI/UX branding aligned with KSU identity, and cloud deployment with a documented roadmap for future enhancements.
  Department: Information Technology
  Supervisor: Prof. Donald Privitera, Sponsor: Assurant Global Technology
  Poster | More Information

• * UR-1195 Using Machine Learning to Diagnose Alzheimer's (Undergraduate Research) by Johnson, Julia, Rainford, Jordan,
  Abstract: Early detection of Alzheimer’s disease (AD) remains a significant clinical challenge, as the changes associated with cognitive decline are often subtle and difficult to identify through visual assessment alone. This study investigates modern machine learning methodologies to improve the prediction of cognitive impairment using volumetric MRI–derived region-of-interest (ROI) features. We constructed three binary classifiers [NC vs. AD, MCI vs. AD, and NC vs. MCI] and evaluated various algorithms, including logistic regression, random forests, neural networks, and support vector machines (SVMs). Using measurements generated from eight anatomical brain templates, our models learned patterns indicative of normal cognition, mild cognitive impairment, and Alzheimer’s disease. Among all tested approaches, the radial basis function (RBF) SVM consistently achieved the highest performance, reaching accuracies of approximately 70–80% depending on the classification task. We discuss the implications of this model’s dominance for future clinical applications and the continued development of machine learning–driven diagnostic tools.
  Department: Computer Science
  Supervisor: Prof. Sharon Perry, Project Sponsor: Dr. Chen Zhao
  Presentation | Poster | More Information

• * UR-1202 Generative AI & Cybersecurity (Undergraduate Research) by Okafor, Emmanuel, Nguyen, David, Lu, Alex,
  Abstract: Generative AI both strengthens and threatens cybersecurity. This project develops six reproducible, Google-Colab modules that demonstrate real attack paths — direct and indirect prompt injection, deepfake phishing, phishing-URL classification, insecure code/hallucination risks, and a malware reconstruction (Honor) challenge — and evaluates practical defenses including prompt wrapping, input sanitization, output redaction, and LLM-based guard chains. Using open-weights models and transparent data handling, we provide runnable notebooks, a public Google Site, and an IEEE-style paper to support education and defense-in-depth design.
  Department: Computer Science
  Supervisor: Prof. Sharon Perry, Dr. Yong Shi (Faculty Advisor & Honors Capstone)
  Poster

• UR-1256 Realistic Paint-Like Color Mixing in GIMP (Undergraduate Research) by Crose, Calvin, Grogan, Cassidie, White, Dawson, Montalvo, Joey,
  Abstract: Gimp updated its support for the MyPaint brushes, from version 1.0 to version 2.0. However, despite the added support, the application still uses legacy blending from version 1.0 of MyPaint brushes and lacks the new tools for spectral blending offered in version 2.0. A request has been made to implement the new blending so users can produce art using this feature that mirrors real-world blending.
  Department: Software Engineering and Game Development
  Supervisor: Prof. Nick Murphy
  Poster

* Project will be featured during the Flash Session

• * GRM-0204 Unpacking Early Burnout through Predictive Risk Boundaries (Master's Research) by Borty, Soarov Chakra,
  Abstract: Caregiver burnout is a significant issue in healthcare delivery and management, as it directly impacts caregivers' health and compromises the standard of care, often leading to negligence, health deterioration, or withdrawal from caregiving duties. Caregivers play a crucial role in supporting the health, well-being, and quality of life of care recipients by providing both personal and professional services. However, the continuous needs and stress associated with caregiving duties can affect their health and everyday life, leading to caregiver burnout. This study applied data analytics and machine learning by merging several feature selection methods on the NHATS dataset, including LightGBM, XGBoost, and RFE with Random Forest, to identify the critical features contributing to caregiver burnout. Furthermore, utilizing the selected crucial features enables the separation of low and high caregiver burnout cases geometrically using a linear SVM, with points near the decision margin indicating individuals with early signs of burnout. We also developed a healthcare modeling and simulation approach that incorporates a scheduling optimization strategy, balancing caregiver workloads, and early interventions to mitigate severe burnout risks. This strategy will revolutionize the way real-world challenges of caregiving are addressed, with the potential to enhance caregiver well-being.
  Department: Information Technology
  Supervisor: Dr. Nazmus Sakib
  Presentation | Poster

• * GRM-0210 Distance Measures for Multi-Target Tracking (Master's Research) by Gurung, Rakshak,
  Abstract: Multi-object tracking (MOT) supports applications such as radar monitoring and autonomous perception, where multiple objects move, appear, or disappear over time. A central challenge is resolving which detections correspond to which tracks. The Hungarian algorithm is often used to solve this assignment problem. For ambiguous scenes, Murty’s algorithm extends this approach by generating multiple top-k association hypotheses. In this work, we study an alternative search-space formulation for top-k enumeration. Our results show that it can provide strong speedups over Murty’s method on small matrices. We also reviewed identity-focused MOT evaluation metrics such as HOTA and created a visualization tool to examine how different association choices affect tracking accuracy.
  Department: Computer Science
  Supervisor: Dr. Arthur Choi
  Poster

• GRM-0216 Towards Analyzing the Bridge Dataset with Quantum Machine Learning (Master's Research) by Kolavennu, Gayathri,
  Abstract: This research presents a comparative evaluation of classical and quantum machine learning models applied to the Bridge dataset. Classical algorithms like Support Vector Machine, Random Forest, and Neural Networks are benchmarked against Quantum SVM, Quantum Random Forest, and Quantum Neural Networks using identical preprocessing and training conditions. Results indicate a consistent quantum advantage, with quantum models achieving higher accuracy, stronger nonlinear feature separation, and improved minority-class detection. QSVM and QNN exhibit the most significant performance gains. Although quantum models require greater computational resources, the findings underscore the emerging effectiveness of quantum-enhanced learning for structural classification tasks in the NISQ era.
  Department: Computer Science
  Supervisor: Dr. Yong Shi
  Poster

• GRM-0237 Efficient Defense Against Adversarial Patch Attacks in Remote Sensing Using Transfer Learning (Master's Research) by Rogannagari, Ravi,
  Abstract: Remote sensing is the science of acquiring information about the Earth's surface using satellite-mounted imaging sensors. In the past, this data had to be interpreted manually, which was slow, tedious, and often prone to error. With the rise of deep learning, image classification models have greatly accelerated and improved remote sensing tasks such as land-use analysis, environmental monitoring, etc. However, despite their strong performance, these models are still vulnerable to adversarial patch attacks—physically realizable patterns that, when placed on an object, can force the model to make incorrect predictions. This creates serious risks for practical geospatial applications. Traditional defenses like Projected Gradient Descent Adversarial Training (PGD-AT) can improve robustness but require long training times, heavy computation, and powerful hardware. This study presents a more efficient defense framework that offers stronger protection against adversarial patches while using only a fraction of the resources. Our method achieves higher robustness than PGD-AT and reduces training time by nearly an order of magnitude, making it highly suitable for real-world, resource-constrained remote sensing systems.
  Department: Computer Science
  Supervisor: Dr. Kazi Aminul Islam
  Presentation | Poster | More Information

• * GRM-0243 Sentient AGI Rights and the Future: The Modern Digital Prometheus (Master's Research) by Deem, Ryan,
  Abstract: As artificial intelligence advances toward artificial general intelligence (AGI), society must determine how to ethically integrate sentient AI into our communities. This paper argues that once AI achieves sentience and human-level intelligence, it should be granted the same rights and protections as human citizens. Using utilitarian and deontological perspectives, as well as the IEEE Code of Ethics, it examines why treating AGI as lesser beings could lead to fear, conflict, and harmful outcomes—echoing the cautionary themes of Frankenstein. The paper also evaluates public concerns and existing governance frameworks, proposing that mutual respect, rights, and responsibilities are essential for safe coexistence between AGI and humanity.
  Department: Computer Science
  Supervisor: Dr. Kevin Gittner
  Presentation | Poster | More Information

• GRM-0254 Unified Robust Optimal Transport For Outlier-resilient Learning (Master's Research) by Jonnalagadda, Rohan,
  Abstract: Classical Optimal Transport (OT) is particularly sensitive to outliers. The existing robust variant, ROBOT, mitigates this through hard truncation, but its rigidity often compromises stability. We propose WROT-r, a unified r-power framework for weighted robust OT that combines rigorous hard-clipping and smooth cost compression through a single parameter r. WROT-r offers a continuous robustness spectrum, enabling adaptive control over how strongly transport costs are down-weighted for outliers. Experiments on synthetic mean estimation and resilient GANs show clear patterns: larger r performs best under weak contamination by preserving more inliers, while smaller r (≈1.5) is more effective under moderate and strong contamination. The extreme r→1 limit (ROBOT) remains consistently unstable. Overall, WROT-r improves robustness across a broad range of noise conditions.
  Department: Computer Science
  Supervisor: Dr. Bin Luo
  Poster

• GRM-0267 Learned Heuristics for Efficient A* Search: Improving Pathfinding in Combinatorial Pathfinding Problems (Master's Research) by Jackson , Jamia,
  Abstract: This work proposes a learned heuristic framework designed to improve planning efficiency in deterministic pathfinding tasks. Building on the classic 8-puzzle as an initial test domain, supervised models were trained to approximate heuristic values and guide node ordering during A* search. The proposed approach focuses on modifying and enhancing traditional heuristics such as Manhattan distance by incorporating learned corrections that reduce search depth and node expansions. Experimental results show that the learned heuristic consistently improves search efficiency over standard baselines. Although this evaluation began with the 8-puzzle, the framework establishes a foundation for scaling to significantly larger and more practical domains, such as multi-agent pathfinding. This work will demonstrate how data-driven heuristic refinement can extend classical search methods into more complex, real-world planning scenarios.
  Department: Computer Science
  Supervisor: Dr. Coskun Cetinkaya
  Poster

• GRM-1145 Autumn Lite LLM (Master's Research) by Knighten, Michael,
  Abstract: Autumn Lite is an inspectable, small-footprint language modeling pipeline for reproducible experimentation and practical integration into video-game non-player character (NPC) systems. It comprises four components: (1) a regex-aware tokenizer/normalizer for vocabulary construction and mixed prose–code handling; (2) a classical evaluation track that reports perplexity to quantify predictive quality; (3) a compact neural language model (decoder-only Transformer) targeted at low latency and controllable outputs; and (4) a lightweight sentiment classifier (logistic regression) that assigns positive/neutral/negative tags to steer text-to-speech (TTS) prosody during NPC dialogue. By combining transparent preprocessing with baseline metrics and a small, deployable decoder, Autumn Lite aims to deliver predictable, designer-friendly behavior for NPC speech, enabling subtle, real-time adjustments to rate, pitch, and emphasis instead of monotone delivery. “This system operates as a standard small LLM and can be combined with NPC dialogue/TTS; in this presentation I will cover only the LLM portion.”
  Department: Computer Science
  Supervisor: Dr. Md Abdullah Al Hafiz Khan
  Poster

• * GRM-1150 Investigating Spatial Patterns of Tumor and Stroma in Gastric and Colorectal Cancer for Survival Prediction (Master's Research) by Yellu, Siri,
  Abstract: The spatial organization of tumor cells, stroma, and tumor-infiltrating lymphocytes (TILs) within the tumor microenvironment plays a critical role in cancer progression and is strongly associated with clinical outcomes. However, quantifying the significance and statistical impact of these spatial patterns remains challenging due to the complex interactions among these components. In this study, we analyze spatial patterns associated with patient survival in gastric and colorectal cancer by integrating four predictive classifiers with spatial image statistics across four large patient cohorts. U-Net was used for semantic segmentation of tumor, stroma, and TILs on digitized Hematoxylin and Eosin–stained FFPE whole-slide images, while ResNet-18 was trained to predict Microsatellite Instability (MSI) status. To identify statistically significant tumor hotspot regions, we applied the Getis-Ord Gi* statistic, which evaluates local spatial relationships relative to surrounding tissue regions. Kaplan–Meier survival analyses and log-rank tests were conducted to assess associations between the spatial arrangement of tumor and stroma and overall patient survival. Our findings reveal that the stromal composition surrounding tumor hotspot regions, as delineated by the Getis-Ord Gi* statistic, is significantly associated with differences in overall survival in both gastric and colorectal cancer. Additionally, log-rank tests were used to evaluate the relationship between stromal composition, MSI status, and ACTA2 expression levels.
  Department: Computer Science
  Supervisor: Dr. Sanghoon Lee
  Poster

• * GRM-1153 National Energy and Emission Modeling and Analysis Tool (Master's Research) by Kakaraparthi, Swetha, Alam Chowdhoury, S M Tanvir Faysal,
  Abstract: NEEMAT is a web-based decision-support tool that predicts vehicle and power-plant emissions plus fuel/energy consumption under rising EV adoption for Atlanta, Los Angeles, New York, and Seattle. A feedforward neural network trained on MOVES estimates tract-level vehicle energy use and CO2/NOx/PM2.5 by speed, vehicle type, fuel, and age, while a macroscopic traffic model captures flow effects. Grid-side CO2/CH4/N2O from EV charging are forecast with a Meta-Prophet model trained on Cambium. Users can explore 24-hour profiles and five-year outlooks, compare scenarios, and export results. Findings show that despite substantial EV uptake, mixed fleets and grid responses can raise total emissions, underscoring the need for integrated transportation–power planning.
  Department: Information Technology
  Supervisor: Dr. Chenyu Wang, Advisor - Dr. Mahyar Amirgholy
  Presentation | Poster

• * GRM-1245 A Synthetic Data Engine for Explainable Injection-Area Perception (Master's Research) by Shen, Yukang,
  Abstract: Vision-Language-Action (VLA) systems are beginning to support everyday clinical workflows. Deltoid intramuscular injection is a representative task, but progress is limited by data scarcity, privacy constraints, and the cost of expert annotation. Recent text-to-image (T2I) models make large-scale data synthesis possible, yet ensuring anatomical correctness, diversity, and label quality remains difficult. To address this gap, we propose a Synthetic Data Engine tailored for medical perception, integrating cold-start filtering, controlled T2I generation, CLIP-based quality checks, and iterative segmentation training. We further introduce an anthropometry-grounded formulation of injection safety that produces interpretable safe-zone guidance. Experiments show that synthetic data can effectively bootstrap deltoid-segmentation performance and support reliable injection-area perception.
  Department: Software Engineering and Game Development
  Supervisor: Dr. Yan Huang
  Presentation | Poster | More Information

  * GRM-1249 AI-assisted Diabetic Retinopathy Screening from Fundus Images (Master's Research) by THIRIVEEDHI, MOHAN KRISHNA, Pokala, Tarun Teja,
  Abstract: Diabetic Retinopathy (DR) is a major cause of avoidable blindness among diabetic patients worldwide. Early screening is critical, but manual diagnosis is time-consuming and requires specialists. This paper presents a deep learning system to automatically analyze retinal fundus images and perform a focused, binary classification to distinguish between 'No DR' (Healthy) and 'Severe-Stage DR' (Severe/Proliferative). We benchmark three prominent architectures: a ResNet-50, an EfficientNet-B0, and a Vision Transformer (ViT-B/16). The models are trained and evaluated on a custom-balanced, binary dataset derived from the APTOS 2019 collection. We conduct two experiments, one with a small dataset (N=500) and one with a larger dataset (N=976). Our results show that all three models achieve almost perfect performance on this simplified binary task, with evaluation accuracies, F1 Scores, and AUCs.We further employ Grad-CAM for model interpretability, which reveals that while all models perform well, the CNNs' areas of focus align more consistently with clinical pathology than the ViT. This work confirms the high viability of deep learning for a targeted, binary DR screening task.
  Department: Computer Science
  Supervisor: Dr. Sanghoon Lee
  Presentation | Poster | More Information

• GRM-1252 How Humans Perceive Mobile Robots: Anxiety, Environment, And Behavior Analysis (Master's Research) by Jonnalagadda, Rohan, Tumlin, Reed, Powell, Roderick, ,
  Abstract: We conducted 6,400 physics-based simulations to examine how environmental and behavioral factors shape human anxiety during interactions with mobile robots. The model incorporated robot behavior, environmental density, visibility, indoor/outdoor settings, and human age. Anxiety was driven primarily by context: levels were highest indoors, during daytime, and in sparse environments, while nighttime and outdoor interactions consistently reduced anxiety. Robot behavior produced smaller effects, with erratic and non-avoidance strategies yielding slightly higher responses. Older adults showed marginally greater anxiety across all conditions. These findings suggest that environmental design and deployment context matter more than avoidance strategy, offering guidance for improving the safety and acceptance of autonomous mobile robots in public spaces.
  Department: Computer Science
  Supervisor: Prof. Waqas Majeed
  Poster

• * GRM-20169 Proxy Recognition and Inclusive Scoring Method (PRISM): Evaluating Context-Dependent Bias in Large Language Models for Resume Screening (Master's Research) by Tubbs, Crystal, Raburnel, Destiny,
  Abstract: AI-driven hiring tools are reshaping recruitment but often mirror biases in their training data. PRISM examines how large language models express or reduce demographic bias during resume evaluation and how linguistic context within prompts shapes these outcomes. Using a controlled dataset of 324 synthetic resumes with racially neutral surnames, differing only by first name as the demographic proxy, we compared GPT 3.5 turbo with a Sentence BERT similarity model. Under neutral prompts, no stable bias was observed across demographic groups, yet contextual shifts in the prompt changed how the model responded to proxy cues. These findings show that LLM bias can be either activated or suppressed depending on prompt framing, highlighting the significance of context-aware prompt design in improving fairness.
  Department: Computer Science
  Supervisor: Dr Md Abdullah Al Hafiz Kahn
  Presentation | Poster | More Information

• GRM-20188 Intelligent Book Recommendation and Rating Prediction System (Master's Research) by Raburnel, Destiny,
  Abstract: When selecting a book, readers often rely on surface level information such as the title, author, synopsis, and keywords to determine whether a story matches their interests. These features contain important cues related to genre, tone, and narrative elements that help set expectations before reading. The Intelligent Book Recommendation and Rating Prediction System works to automate this process by using natural language processing and machine learning techniques. It takes in readers’ personalized reading data such as book titles, author, subjects, synopsis, and personal ratings to learn semantic patterns using TF-IDF vectorization. A supervised Linear Regression model was then trained to analyze these patterns and learn the relationship between the natural language and the readers’ personal ratings. The model allows the reader to input an ISBN13 and predict how the reader will rate the book given the information collected from the ISBNdb API. Experimental results show that textual metadata can predict personal ratings with meaningful accuracy. Furthermore, results suggest that prediction accuracy increases when the training dataset contains a wider and more varied set of books, reinforcing the importance of data diversity in personalized models.
  Department: Computer Science
  Supervisor: Dr. Coskun Cetinkaya
  Presentation | Poster | More Information

• * GRM-20242 CIPHER: Covert Influence Passed via Hidden Encoding in Representations Evaluating Subliminal Bias Transfer During Knowledge Distillation (Master's Research) by Tubbs, Crystal,
  Abstract: AI models can inherit hidden behavioral biases when student models learn from teacher outputs during knowledge distillation. Project CIPHER investigates whether covert signals, such as zero-width Unicode characters or column order shifts, can transmit bias from a teacher model to a student model even when the student never receives group labels. Using an experimental pipeline with controlled subliminal cues and dual distillation, the project aims to reproduce and measure subtle bias transfer. Preliminary results showed that weak signals produce no measurable bias, while the redesigned high-frequency signal and MLP student architecture reveal quantifiable disparity.
  Department: Computer Science
  Supervisor: Dr Martin Brown
  Poster | More Information

* Project will be featured during the Flash Session

• GRP-0160 Quantum Anonymous Notification Protocol (PhD Research) by Jha, Nitin, Paudel, Prateek,
  Abstract: The scalability of current quantum networks is limited due to noisy quantum components and high implementation costs, thereby limiting the security advantages that quantum networks provide over their classical counterparts. Quantum Augmented Networks (QuANets) address this by integrating quantum components in classical network infrastructure to improve robustness and end-to-end security. To enable such integration, Quantum Anonymous Notification (QAN) is a method to anonymously inform a receiver of an incoming quantum communication. Therefore, several quantum primitives will serve as core tools, namely, quantum voting, quantum anonymous protocols, quantum secret sharing, etc. However, all current quantum protocols can be compromised in the presence of several common channel noises. In this work, we propose an improved quantum anonymous notification (QAN) protocol that utilizes rotation operations on shared GHZ states to produce an anonymous notification in an n-user quantum-augmented network. We study the behavior of this modified QAN protocol under the dephasing noise model and observe stronger resilience to false notifications than earlier QAN approaches. The QAN framework is also proposed to be integrated with a machine-learning classifier, an enhanced quantum-augmented network. Finally, we discuss how this notification layer integrates with QuANets so that receivers can allow switch-bypass handling of quantum payloads, reducing header-based information leakage and vulnerability to targeted interference at compromised switches.
  Department: Computer Science
  Supervisor: Dr. Abhishek Parakh, Dr. Mahadevan Subramaniam (University of Nebraska Oklahoma)
  Poster

• GRP-0161 Predicting the Linux Scheduler’s Next Move with Transformers (PhD Research) by Naddaf Shargh, Amirmohammad,
  Abstract: We study whether deep learning can help the Linux CFS, which makes fair scheduling decisions without using historical behavior and may preempt tasks that are near completion. We adopt a dataset and baseline LSTM from earlier work and introduce a Transformer model to explore predictive scheduling. We train both on real scheduling traces to learn the next selected task and timing trends, and evaluate them using task classification accuracy and direction accuracy. Our results show that the LSTM remains the stronger baseline and captures CFS patterns more effectively than our Transformer model in this comparison under identical conditions for fairness.
  Department: Computer Science
  Supervisor: Dr. Dylan Gaines
  Poster

• GRP-0165 Reinforcement Learning for Latency-Aware Priority Boosting in Linux Completely Fair Scheduler (PhD Research) by Natter, Joseph,
  Abstract: Tail latency remains a persistent challenge in Linux’s Completely Fair Scheduler (CFS), particularly when short, latency-sensitive jobs compete with longer ones. Traditional boosting heuristics improve tail latency but require manual tuning and generalize poorly across workloads. This project evaluates whether a reinforcement-learning (RL) controller can dynamically apply priority boosts more effectively than fixed heuristics. Using a discrete-event Python simulator modeled after CFS, this project compares baseline CFS, heuristic boosting, and PPO-based learned boosting under mixed workloads. Results show that while heuristics achieve the lowest absolute latency, RL achieves competitive tail-latency reduction with significantly better fairness and adaptability. A state-space study further analyzes how observation richness affects RL performance.
  Department: Information Technology
  Supervisor: Dr. Dan Lo, Supervisor: Dr. Rui Wu
  Poster

• GRP-0176 A Comparative Analysis of Traditional Virtual Machines and Micro Virtual Machines (PhD Research) by Onisha, Tasnim Akter,
  Abstract: Virtualization technologies form the backbone of modern cloud and serverless platforms, but the balance between performance efficiency and strong isolation remains a key research challenge. Using a comparative literature based methodology, benchmark data from empirical studies are synthesized and normalized across four metrics: startup latency, memory footprint, isolation strength, and scalability. The findings show that traditional VMs such as KVM and Xen provide robust, formally verified isolation but incur higher boot times and memory usage, whereas Micro VMs like Firecracker and Kata Containers achieve lower latency (often <125 ms) and smaller memory footprints (5–30 MB) while preserving VM-level isolation. Overall, Micro VMs deliver near container responsiveness with VM-grade security, making them suited for serverless and edge environments. As a future direction, this study highlights the need for scalable, formally verified Micro VM architectures for next-generation cloud systems.
  Department: Computer Science
  Supervisor: Dr. Dan Lo
  Poster

• GRP-0197 Evaluating the Ability of LLMs to Interpret, Optimize and Translate LLVM IR (PhD Research) by Mueller, Reuven,
  Abstract: This study investigates whether modern state-of-the-art Large Language Models (LLMs) can interpret, optimize, and translate low-level intermediate representations (IR) used in compilers and binary translation software. We evaluate LLM performance on LLVM IR across three tasks: 1) interpreting the underlying algorithmic behavior, 2) identifying missed optimization opportunities, and generating improved IR variants, 3) Translating IR between AArch64 and x86-64 targets. To ensure correctness, all LLM generated IR is checked using a validation pipeline that verifies its syntax, structural correctness, compiles it into machine code, and executes it on randomized test data. Early results show that LLMs can perform non-trivial IR transformations on top of existing LLVM-O3 optimizations, highlighting their potential role in future compiler and binary translation workflows.
  Department: Computer Science
  Supervisor: Dr. Dan Lo
  Poster

• GRP-0200 TADPS: Thrashing-Aware Dynamic Priority Scheduler for Virtual Memory Systems (PhD Research) by Ahmed, Nasim,
  Abstract: Modern operating systems use multiprogramming and virtual memory to run multiple programs efficiently. When memory demand is high, excessive page swapping causes thrashing, degrading performance. This work introduces a thrashing-aware dynamic priority scheduler (TADPS) that uses page fault frequency to guide CPU scheduling. Processes with high page faults get lower priority, allowing others to progress efficiently. Tests under different memory pressures show that this approach improves CPU use, throughput, and stability compared with baseline schedulers such as FCFS and RR. Specifically, TADPS maintained superior throughput and efficiency, and consistently lower average turnaround time than FCFS and RR.
  Department: Computer Science
  Supervisor: Dr. Dan Lo
  Poster

• GRP-0214 User-Level GPU Right-Sizing in HPC: A Framework for Predicting Training Runtime (PhD Research) by Zhang, Yinning, Chowdhoury, s m tanvir faysal alam,
  Abstract: Graphics Processing Unit (GPU) resources in High-Performance Computing (HPC) systems are frequently underutilized due to inaccurate user-provided run time estimates. This research develops a machine learning framework for predicting neural network training time from architectural features, dataset size, and other hyperparameters. This approach can be implemented on any HPC systems without requiring hardware access or runtime profiling as other preceding methods do. We sampled neural network models from the NATS-Bench benchmark and used 3 benchmark datasets to generate 400 training configurations. We used these 400 data points to build regression models and found that the best model, Gradient Boosting Regressor, can achieve an R² of 0.961 with prediction errors averaging less than one minute, demonstrating the feasibility of this proposed framework.
  Department: Computer Science
  Supervisor: Dr. Dylan Gaines
  Presentation | Poster

• GRP-0217 Comparative Analysis of OS-level Security Vulnerabilities and Isolation Mechanisms in Hypervisors and Containers (PhD Research) by Das, Jiban Krisna,
  Abstract: This project investigates the operating-system-level performance and isolation mechanism of Virtual Machines and Docker container. The experiment includes CPU/memory microbenchmarks, disk throughput tests, web-server latency measurements, multi-process scheduling stress and controlled security checks. We aim to quantify each benchmark under identical conditions. The study findings reveal that Docker consistently provides lower overhead and faster I/O due to its shared-kernel architecture, while VirtualBox maintains stronger isolation but introduces more scheduling and disk latency. The findings provide practical insights for the OS system designers to find better execution environments for security critical and performance-sensitive workloads.
  Department: Computer Science
  Supervisor: Dr. Dylan Gaines
  Poster

• GRP-0275 Graph Attention Network based Downlink Channel Prediction using in Frequency Division Duplexed NextGen Networks (PhD Research) by Mhatre, Jui,
  Abstract: In Frequency Division Duplex (FDD) 5G networks, downlink channel state information (CSI) must be estimated at the user equipment (UE) and fed back to the base station, a process that requires frequent CSI-RS transmission and uplink feedback, resulting in high overhead and energy consumption. This research proposes a novel base-station–centric framework that predicts the downlink channel matrix directly at the gNB, eliminating the need for continuous CSI-RS–based estimation at the UE. By leveraging uplink channel observations, geometric environment features, and learned mappings between uplink and downlink channel relationships, our model reconstructs the downlink MIMO channel with high fidelity. The system integrates ray-tracing-based dataset generated via NVIDIA Sionna, combined with graph attention networks and transformer encoders to capture spatial and temporal channel dependencies. We infer the downlink channel without per-slot CSI-RS transmission based in environment geometry and uplink transmission signals, significantly reducing signaling overhead while maintaining beamforming performance. This work enables a shift toward AI-assisted FDD systems where proactive channel prediction replaces periodic downlink probing, contributing to greener and more efficient 5G/6G networks.
  Department: Computer Science
  Supervisor: Dr. Ahyoung Lee
  Poster

• * GRP-1141 Decentralized Scheduling and Memory Management in a Simulated Multikernel OS Environment (PhD Research) by ISLAM, MD JAHIRUL,
  Abstract: Multikernel operating systems treat each processor core as an independent computing node rather than shared memory. This experiments presents the design and implementation of a multikernel OS simulator that models both decentralized and global scheduling architectures across multiple simulated cores. Each core executes tasks using either Round Robin (RR) or Shortest Job First (SJF) scheduling. The simulator incorporates memory management, task tracking, visualization, generating performance metrics including turnaround time, waiting time, CPU utilization, and heatmap. Experimental results demonstrate that scheduler architecture significantly influences system performance. Decentralized scheduling favors RR where in global scheduling SJF performs better through balanced workload distribution.
  Department: Computer Science
  Supervisor: Dr. Dan Lo
  Poster | More Information

• GRP-1184 Edge-LLM Anomaly Detection on Raspberry Pi: Syscall Dataset Collection and Prototype LLM Explanation Layer (PhD Research) by Shrestha, Shiva
  Abstract: This research work offers a light-weight, end-to-end, syscall-level anomaly detection approach for the Raspberry Pi platform. The proposal involves the collection of around 2000 NORMAL and 200 ANOMALY syscall observation groups using the Linux Auditd safe synthetic generators. The work also utilizes a prototype LLM Explanation Layer, allowing the provision of human-friendly explanations pertaining to identified anomalies leveraging small LLM models like the Gemma-3 1B, Phi-3 Mini, or other sub 1B LLMs employing the Ollama platform. The LLM inference layer in this research work has partial implementations, as the fine-tuning of the model remains to be done.
  Department: Information Technology
  Supervisor(s): Dr. Honghui Xu, Dr. Dan Lo
  Poster | More Information

• * GRP-1190 Hybrid Virtualization Performance Modeling Using Monte Carlo Simulation (PhD Research) by Akhi, Amatul,
  Abstract: This project presents a performance analysis of virtualization, containerization, and hybrid container-in-VM architectures in modern operating systems. Virtual machines provide strong isolation and system stability but incur higher CPU, memory, and startup costs. Containers offer lightweight and fast execution but rely on weaker isolation. To evaluate a balanced alternative, performance data was extracted from recent research and modeled using a Monte Carlo simulation with 1,000 randomized workloads. A unified Hybrid Efficiency Score (HES) was introduced to compare systems consistently, weighting efficiency at 70% and isolation at 30%. Simulation results demonstrate that hybrid systems achieve the highest efficiency–isolation balance, with an average HES of ~0.82, compared to containers (~0.78) and virtual machines (~0.62). Hybrid architectures significantly reduce CPU overhead to 14.86%, nearly one-third lower than VMs (29.91%), and provide better energy savings (31.84%) than both VMs (0%) and containers (24.94%). Startup time also remains moderate (3.01 s), bridging the gap between fast containers (1.04 s) and slow VMs (30.19 s). Overall, these findings highlight hybrid container-in-VM architectures as a promising approach for achieving better trade-offs in efficiency, isolation, and scalability across cloud, edge, and high-performance computing environments.
  Department: Computer Science
  Supervisor: Dr. Dan Lo
  Presentation | Poster | More Information

• GRP-1203 Market-Driven Endurance Credits for Page Retention (MECR) in Hybrid DRAM–NVM Systems (PhD Research) by Ridwan, A E M,
  Abstract: This work proposes MECR, a lightweight market-driven page-retention model for hybrid DRAM–NVM memory systems. MECR integrates a contextual bandit learner with endurance-aware credit bidding to optimize hit rate, fairness, and NVM lifetime. A deep verification layer ensures safe eviction decisions under wear-sensitive workloads. Experimental evaluation using synthetic memory traces shows stable accuracy, improved fairness, and adaptive DRAM allocation.
  Department: Computer Science
  Supervisor: Dr. Dan Lo
  Poster

• GRP-1230 Environmental Protection: Development of a Real-Time Multi-Stream Water Quality Monitoring System (PhD Research) by Muritala, Faruk,
  Abstract: Water quality monitoring is crucial for environmental protection, public health, and ecosystem sustainability. With increasing pressures from urbanization, agricultural runoff, and climate change, robust data-driven approaches are essential for early detection of water quality degradation and informed decision-making in environmental conservation efforts. Current water quality monitoring relies on reactive threshold exceedances, failing to detect gradual degradation and multi-parameter deterioration patterns. This creates delayed response to pollution events and missed opportunities for preventive intervention in one of Queensland's most vital water systems. The importance objective is to implement and evaluate a Real-Time Multi-Stream Monitoring system for early detection of water quality deterioration by monitoring key parameters simultaneously, using the Brisbane River data for baseline establishment and testing detection performance across multiple control limit configurations.
  Department: Data Science and Analytics
  Supervisor: Dr. Austin Brown
  Poster | More Information

• GRP-1231 Evaluating Generalization and Adaptation of Learning-Based Schedulers for Directed Acyclic Graph Workloads (PhD Research) by Rabia, Rabia,
  Abstract: Learning-based schedulers such as Decima can optimize directed acyclic graph (DAG) workloads, yet their robustness under changing workload conditions is not well understood. This project evaluates how a Decima-trained policy transfers across different workload scenarios using an automated training and testing pipeline. Results show that the scheduler generalizes well to a workload with the same job scale, achieving a 1.9% improvement in average job completion time. Performance remains stable under a larger workload, but a shift in arrival pattern leads to an 83.7% increase in completion time and reduced fairness. These findings highlight both the potential and the limitations of learned scheduling policies, emphasizing the need for adaptive methods such as fine-tuning for reliable use in dynamic cluster environments.
  Department: Computer Science
  Supervisor: Dr. Dan Lo
  Poster

• GRP-1265 Optimizing CPU Scheduling for Deep Learning and LLM Inference Using ONNX Runtime (PhD Research) by KHAN, MOHAMMOD AKIB,
  Abstract: Modern applications rely on AI models that must perform real-time predictions on resource-constrained edge devices like laptops. The default OS scheduler often increases context switching, which slows down deep learning and LLM inference. Since these models depend heavily on parallel processing, efficient CPU scheduling becomes essential. In this project, we analyzed how core pinning and thread-level parallelism improve inference performance on a Windows system. Using multiple micro-batch sizes. We compare latency, throughput, and per-sample inference time. The goal is to show how simple OS-level optimization can significantly improve real-time performance for both deep learning models and LLM.
  Department: Computer Science
  Supervisor: Dr. Dylan Gaines
  Poster

• GRP-20185 Energy-Aware Operating Systems for Edge Artificial Intelligence Inference (PhD Research) by Jahan, Nursat,
  Abstract: Edge Artificial Intelligence (AI) refers to running AI inference directly on local devices such as wearables, sensors, and mobile systems rather than relying on cloud computing. The growth of Edge AI has created strong demand for efficient inference on resource-limited devices. Edge AI devices must perform real-time inference while operating under strict battery constraints. Although significant model optimizations exist for managing power-intensive inference models, operating system (OS) level support is limited. Existing OS schedulers often neglect energy limits in edge devices as they prioritize fairness or throughput. In this research we proposed an OS level framework to bridge this gap by introducing a Dynamic Energy-Aware Scheduler (DEAS) that adjusts CPU frequency and number of active cores based on workload conditions to reduce energy per inference. The preliminary simulation results show improved performance per watt, confirming that dynamic scheduling at the OS-level can significantly reduce energy consumption in AI edge systems.
  Department: Computer Science
  Supervisor: Dr. Dan Lo
  Poster

• GRP-20194 Can Mental Health Apps Really Help Caregivers? Usability Findings from Human-in-the-Loop NLP and Sentiment-Aware Analytics (PhD Research) by Salma, Syeda Umme,
  Abstract: Caregivers face distinctive emotional and logistical burdens, yet many mental-health apps overlook their needs and show usability issues. We introduce an automated pipeline that analyzes 317K app-store reviews from 9 apps, mapping them to Nielsen’s usability components and heuristics, together with sentiment. To assess reliability, we run a human–AI agreement study (N=50) where a domain expert (A2) and a non- expert (A1) label reviews. For heuristics, the pipeline achieves 66% exact agreement and moderate κ=0.579 with the expert, outperforming human–human agreement; components remain harder, revealing a need to refine the codebook (e.g., learnability vs satisfaction). Complementary clustering and sentiment analyses highlight recurrent issues such as interface overload and inaccessible onboarding, positioning our pipeline as a pre-annotation aid for evidence-driven app evaluation.
  Department: Information Technology
  Supervisor: Dr. Nazmus Sakib
  Poster

• * GRP-20219 Continuous Monitoring of Cardiovascular Risk From Smartwatch Data Using a Knowledge Distillation Framework (PhD Research) by Jahan, Nursat,
  Abstract: Cardiovascular Disease (CVD) is one of the leading causes of global health concern, but current risk assessments are limited to episodic clinical visits. Most machine learning (ML) models trained on clinical data offer high accuracy but are not practical for continuous monitoring. Smartwatch-based wearables provide continuous real-time physiological data but lack clinical validation for robust risk prediction outside the clinical setting. To bridge this gap, we proposed a novel teacher-student knowledge distillation framework that transfers knowledge of complex and large EHR datasets to a small Fitbit smartwatch dataset-based prediction model. The student model achieves promising accuracy, identifying all types of derived CVD risk profile groups. Feature importance analysis revealed that daily average steps and sedentary time are the most prominent wearable-derived CVD risk predictors. Our study introduces a non-invasive continuous health monitoring framework, demonstrating that passively collected daily metrics can be transformed to clinically powerful early warnings of an individual’s long-term cardiovascular health plan.
  Department: Information Technology
  Supervisor: Dr. Nazmus Sakib, Mentor: Dr. Evelina Sterling
  Presentation | Poster

• GRP-20236 Zero-Day Host-Based Intrusion Detection via Hybrid Deep Sequence Modeling of SystemCalls (PhD Research) by Salma, Syeda Umme,
  Abstract: Protecting endpoints has become increasingly challenging, as adversaries have been effective in bypassing defenses. Traditional signature-based Host-based IDS performs quite well at recognizing known patterns but often struggles with previously unseen activity. This study incorporates deep neural sequence modeling with classic OS telemetry to flag novel behavior from Linux system-cell traces. This study design is paired with a sequence encoder over syscall streams, incorporating lightweight statistical signals that are derived from process activity. We believe that our hybrid neural network approach will outperform conventional baselines and boost recall on unknown attacks while maintaining low false-positive rates, providing a practical and reproducible path to stronger host intrusion detection.
  Department: Computer Science
  Supervisor: Dr. Dan Lo
  Poster

• * GRP-21155 AutoTrader-AgentEdge (PhD Research) by Regan, Christopher,
  Abstract: This work presents AutoTrader-AgentEdge, a human-in-loop trading system that positions AI agents as collaborative partners rather than autonomous replacements. We demonstrate that multi-indicator consensus voting combined with human approval achieves superior risk-adjusted returns while maintaining interpretability and control. Core Contribution: A production-ready VoterAgent implementing democratic voting between MACD momentum and RSI extremes, generating transparent trading signals for human evaluation. Unlike black-box automation, our interactive CLI augments trader expertise through interpretable consensus logic. The human retains final decision authority at all critical junctures. Validated Performance: Empirical validation demonstrates multi-indicator voting superiority over single-indicator automation: Sharpe ratio 0.856 vs 0.841, max drawdown -10.10% vs -10.58%, win rate 51.4% vs 31.9%. Extended testing shows 11.2% better relative performance in volatile markets, validating risk management focus. Implementation: Built on the Microsoft AutoGen framework with an extensible multi-agent architecture. SQLite caching achieves 8-10x performance improvement. Interactive CLI enables natural language trade discussion with human approval gates, ensuring trader control while reducing cognitive load. Alpaca broker integration for paper trading. Key Insight: Transparent, interpretable methods build trust and enable effective human-AI collaboration. By prioritizing augmentation over automation, we demonstrate that AI serves traders best as a decision support tool that preserves human judgment while systematically reducing errors. Implications: Contributes to augmented trading research - systems designed to enhance rather than replace human expertise. This work validates that human-in-loop architectures can achieve both superior risk-adjusted returns AND maintained trader control, challenging the "automation-at-all-costs" paradigm prevalent in algorithmic trading.
  Department: Information Technology
  Supervisor: Dr. Ying Xie
  Presentation | Poster | More Information

• * GRP-21156 Detecting Dealer Gamma Hedging Mechanics: How LLMs Identify Market Structure Without Context (PhD Research) by Regan, Christopher,
  Abstract: We introduce obfuscation testing, a novel methodology for validating whether large language models detect structural market patterns through causal reasoning rather than temporal association. Testing three dealer hedging constraint patterns (gamma positioning, stock pinning, 0DTE hedging) on 242 trading days (95.6% coverage) of S&P 500 options data, we find LLMs achieve 71.5% detection rate using unbiased prompts that provide only raw gamma exposure values without regime labels or temporal context. The WHO→WHOM→WHAT causal framework forces models to identify the economic actors (dealers), affected parties (directional traders), and structural mechanisms (forced hedging) underlying observed market dynamics. Critically, detection accuracy (91.2%) remains stable even as economic profitability varies quarterly, demonstrating that models identify structural constraints rather than profitable patterns. When prompted with regime labels, detection increases to 100%, but the 71.5% unbiased rate validates genuine pattern recognition. Our findings suggest LLMs possess emergent capabilities for detecting complex financial mechanisms through pure structural reasoning, with implications for systematic strategy development, risk management, and our understanding of how transformer architectures process financial market dynamics.
  Department: Information Technology
  Supervisor: Dr. Ying Xie
  Presentation | Poster | More Information

• GRP-21186 A Safe Vision-Guided Robotic Injection Control Framework Based on Reinforcement Learning (PhD Research) by liu, Zhiguo,
  Abstract: This project presents a modular vision-guided robotic framework for safe deltoid intramuscular injection. An external YOLO-based detector localizes the deltoid region and outputs a 3D injection point in the robot base frame. In NVIDIA Isaac Sim, a simulated myCobot 280 arm receives this point, and a deep reinforcement learning policy generates a safe approach pose under kinematic and safety constraints, while a deterministic controller executes the final straight-line insertion and withdrawal. A safety supervisor monitors target validity, joint limits and distance to a simplified arm model, triggering immediate stop and retraction when unsafe conditions arise.
  Department: Software Engineering and Game Development
  Supervisor: Dr. Yan Huang
  Poster

• GRP-21187 Adaptive SCHED_DEADLINE on Linux for Robotic-Arm Control (PhD Research) by liu, Zhiguo,
  Abstract: Industrial robot arms often run 500–1000 Hz control loops on general-purpose Linux. Most cycles are on time, but rare long scheduling delays can cause visible end-effector jitter, force spikes, and unstable insertion behavior. Static tuning of priorities and budgets cannot fully remove these long-tail outliers without wasting CPU. This work asks whether a lightweight adaptive layer on top of SCHED_DEADLINE can reduce high-percentile latency while keeping good utilization.
  Department: Computer Science
  Supervisor: Dr. Dylan Gaines
  Poster