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)

Academic courses undergraduate (e.g. capstones, games, innovative special topics projects, other course projects)

• UC-129 Angel Among Us Pet Rescue - Website Enhancement with Chatbot (Undergraduate Project) by , , , 
  Abstract: This project integrates an AI-powered chatbot into the Angels Among Us Pet Rescue website, enhancing user experience by efficiently addressing common queries. The chatbot uses large language model (LLM) technology, which in this project is ChatGPT, to understand and respond to user questions dynamically. A content management system (CMS) supports easy updates to the chatbot鈥檚 responses, allowing Angels Among Us staff to manage FAQ entries without technical intervention. The chatbot integrates seamlessly with the existing website, maintaining the organization鈥檚 aesthetic, accessibility, and compatibility across devices. This enhancement improves user engagement and streamlines support, enabling the nonprofit to focus more on its core mission of animal rescue. By providing accessible and accurate information through an intuitive interface, the chatbot is a pivotal tool in reducing repetitive inquiries and driving user interaction with Angels Among Us Pet Rescue鈥檚 online resources.
  Department: Software Engineering and Game Development
  Supervisor: Dr. Yan Huang
   | 
  
  

• UC-131 Karah Khronicles (Undergraduate Project) by , Stipetich, Jake, Bowe, Grace, , 
  Abstract: Karah is a thief with a heart of gold, you raid enemy camps and dungeons to steal back the money stolen from towns and villages and upgrade enchanted items to deal with dangerous foes. After successfully returning the wealth to the local town, you must then face down and defeat a general of the evil king.
  Department: Software Engineering and Game Development
  Supervisor: Dr. Sungchul Jung
   |  | 
  
  

• UC-133 Biomedical Deep Learning - A staged approach using trustworthy deep learning for multi-omics data classification (Undergraduate Project) by An, Yongbo, Liu, Tianze
  Abstract: Genetic data such as mRNA, miRNA, and DNA methylation offer precious insights into the underlying causes variant diseases. These types of data provide various layers of information, simultaneously enhancing our understanding of the disease and improving diagnostic accuracy. Combining mRNA, miRNA, and DNA methylation data allows for a multi-dimensional approach to identifying biomarkers, potentially leading to earlier and more accurate diagnosis. However, integrating all modalities is not practical. The clinical cost increases significantly with every modality incorporated. In contrast to previous methods, our model uses partial modalities when possible. We will use subjective logic and trustworthy deep learning under the staged approach to perform disease risk prediction. During our research process, we explored effective modality combinations for single view and bi-view models, designed an optimized multi-perception layer architecture for single-view classification, and implemented methods to quantify and optimize uncertainty in incomplete multi-omics data integration.
  Department: Computer Science
  Supervisor: Dr. Chen Zhao
   | 
  
  

• UC-134 Volunteer Management System for Angels Among Us (Undergraduate Project) by , , , , 
  Abstract: This project focused on creating a Volunteer Management System (VMS) for Angels Among Us (AAU) - a non-profit organization dedicated to rescuing and rehabilitating stray and abandoned animals. This application was developed to: * Handle comprehensive volunteer information * Streamline operations and better manage volunteer data * Support AAU specific use cases * Include a data enrichment capability through a newly developed GUI * Allow authorized users to add more comprehensive information to each volunteer record * implement reporting throughout the data migration process
  Department: Information Technology
  Supervisor: Prof. Donald Privitera
   | 
  
  

• UC-140 Streamlining School Bus Monitoring: GCPS's Transition to Real-Time Kafka Event Processing (Undergraduate Project) by Fashinasi, Sarah, , , , 
  Abstract: This project develops a prototype real-time bus monitoring system for Gwinnett County Public Schools using simulated Kafka event streaming to replace current API polling methods. The system processes simulated Asset Location and Speed events, mimicking Samsara's Kafka Connector, performing data validation before storing in SQL Server. The containerized solution demonstrates the potential for near real-time visibility into school bus operations.
  Department: Software Engineering and Game Development
  Supervisor: Dr. Yan Huang, Ed Van Ness - GCPS Technology and Innovation Division (Sponsor)
   | 
  
  

• UC-141 IT Capstone Project 17 - KSU eSports Tournament Bot (Undergraduate Project) by Helfrick , Patricia G, Jayakumar , Niranjanaa, Schroeder, Daniel J, , Stogsdill, Jackson M
  Abstract: In this project, our team has automated tournament tasks in the KSU eSports Discord server, with a focus on the League of Legends tournaments. Our team has implemented a matchmaking algorithm that forms teams consisting of players placed within one tier of each other, so teams are evenly matched. Our team has also created a database that stores player statistics and has been integrated with the Discord bot. Furthermore, our team has integrated the developer API with the Discord bot, which pulls player data from the API when players join the server, and the team has been working to improve the UI of the bot. The team has performed regular testing and will continue to perform regular testing and updates as necessary. Department: Information Technology Supervisor: Prof. Donald Privitera Topics: Programming
  Department: Information Technology
  Supervisor: Prof. Donald Privitera, project sponsor Kylie Nowokunski
   | 
  
  

• UC-144 Attack Surface Management and Analysis (Undergraduate Project) by , , , , ,
  Abstract: Recent advancements in AI have made knowledge more accessible, but this also introduces risks, as vulnerabilities can now be quickly found and exploited. To address this, we developed a comprehensive, cloud-native attack surface monitoring suite in Google Cloud. Integrating open-source intelligence tools like OWASP Amass and Project Discovery, along with custom Python-based processing, we gather extensive security data鈥攃overing subdomain enumeration, open ports, HTTP responses, and DNS configurations. This data is stored in BigQuery, processed, and visualized in Looker Studio for easy client interpretation. A containerized, scalable backend with a Flask-based API ensures seamless tool integration and adaptability. BigQuery ML further classifies domains鈥� security, empowering organizations with proactive risk assessment and attack surface monitoring.
  Department: Computer Science
  Supervisor: Prof. Sharon Perry
   |  | 
  
  

• UC-145 Eerie (Undergraduate Project) by , Whorton, Joshua, , 
  Abstract: Eerie is a psychological horror/thriller game that plunges players into the harrowing journey of Alice, a young girl trapped in her home. As she navigates the dimly lit corridors of her once-familiar environment, Alice grapples with haunting hallucinations and a distorted reality that intertwines the tangible and surreal. The gameplay revolves around her desperate quest to recover cherished belongings, each revealing deeper layers of her fractured story. Players must confront both real enemies and manifestations of Alice鈥檚 psyche, creating a tense dynamic that challenges them to strategize against both physical threats and the shadows of her fears.
  Department: Software Engineering and Game Development
  Supervisor: Prof. Kevin Markley (Spring), Dr. Sungchul Jung (Fall)
   |  | 
  
  

• UC-150 Azure Migration Assistant (Undergraduate Project) by , , , Ngah, Yvan
  Abstract: Migrating to the Azure cloud platform poses unique cost-assessment and planning challenges. Our project introduces a user-friendly, AI-driven tool to simplify this process by providing real-time cost predictions and personalized migration strategies. Built with a React frontend and a Flask-based Python backend, this tool integrates Azure Pricing APIs to ensure accurate data. Future improvements include adding alerts, custom fine-tuned model, CI/CD, multi-cloud support, and a discovery agent for enhanced functionality.
  Department: Computer Science
  Supervisor: Prof. Sharon Perry, Andrew Anderson -- GTRI Sponsor
   |  | 
  
  

• UC-156 SWAP - A Solo Developed FPS Game (Undergraduate Project) by 
  Abstract: SWAP is an FPS game that blends tactical thinking with quick reflexes and player expression. Your dog Chomper has been kidnapped by the Big Dogs Mafia, and you must infiltrate their undercover locations to bring Chomper back home safe and sound. Along the way, the player will be asked to think on the fly, grabbing anything they can get their hands on to use as a weapon. From pistols and shotguns to forks, screwdrivers and keyboards, everything that the player can pick up is a deadly weapon.
  Department: Software Engineering and Game Development
  Supervisor: Prof. Sungchul Jung
   |  | 
  
  

• UC-166 The Eternal Guest - A 2D Hack-and-Slash Game (Undergraduate Project) by , , , , 
  Abstract: The Eternal Guest is a narrative-driven, hack-and-slash combat and exploration game where you make meaningful friendships, battle enemies, and regain lost memories as you traverse a strange, non-euclidian hotel. Use a wide array of weapons and abilities alongside the knowledge you gain from other guests to attempt to escape the bloodlust of a homicidal vampire. The Eternal Guest emphasizes 2D, top-down, melee combat in combination with ranged abilities, offering an exciting dynamic to gameplay. Our game also presents a unique spin on randomized exploration through its unique D.R.E.A.D. system, creating a sense of unease and uncertainty when exploring. This unique combat style and exploration, combined with a compelling narrative, keeps players returning not only for the story but for the gameplay as well.
  Department: Software Engineering and Game Development
  Supervisor: Dr. Sungchul Jung
   |  | 
  
  

• UC-173 Leveraging Large Language Models to Empower Caretakers of People with Dementia (Undergraduate Project) by , 
  Abstract: Behavioral symptoms of Alzheimer's Disease and Related Dementias (ADRD) are detrimental to the quality of life for individuals with ADRD and their caregivers. Symptoms such as wandering, agitation, and confusion can often overwhelm caregivers leading to stress, depression, or burnout which can lead to a decrease in the quality of care. These challenges often result in increased hospitalizations and care costs, creating a need for a solution to support informal caregivers. This project proposes the development of an AI-based Dementia Care Voice Assistant application to meet the needs of caregivers. Using large language models, the application will provide real-time and personalized guidance to help caregivers manage complex behavioral symptoms. The LLMs will be designed to adapt responses to the user based on how they are prompted. To ensure that the output aligns with the best medical practices, we will establish a dataset based on evidence-based interventions from extensive literature reviews and interviews with informal caregivers. In addition to providing tailored responses, the application will offer assistance during emergency situations. The voice assistant will feature intuitive features such as recognizing signs of medical emergencies and prompting the user to contact 911 when necessary. Through the development of this application, informal caregivers will have access to accurate information and personalized assistance, alleviating caregiver stress, enhancing their confidence, and ultimately improving the quality of the care they deliver.
  Department: Computer Science
  Supervisor: Dr. Xinyue Zhang, Dr. Modupe Adewuyi
   | 
  
  

• UC-176 Cybriant: Attack Surface Management (Undergraduate Project) by Rai, Diwakar, Agyen-Frempong, Nicholas, Laurent, David, Gutierrez, Daniel, Mendoza, Jose R
  Abstract: As businesses and organizations expand their operation digitally, so too do the vectors for attack expand. In partnership with Cybriant, this application develops an Attack Surface Composite Score by breaking down various attack common vectors. DKIM records, Open Port Scanning, and other metrics are compiled with the aid of Google Cloud Run jobs, deposited into Google BigQuery for analysis, and packaged and generated using (Grafana/Kibana) as the front-end for our software stack. Our resulting application presents rapid, easy-to-understand breakdowns of various cybersecurity metrics and their impact.
  Department: Software Engineering and Game Development
  Supervisor: Dr. Yan Huang, Prof. Donald Privitera // project Sponsor mentors Byron DeLoach, Pramit Bhatia, Andrew Hamilton, Sean Mitchell
   | 
  
  

• UC-180 Intelligent Object Retrieval using Mobile Manipulator (Undergraduate Project) by Zheng, Zhiwen, 
  Abstract: A mobile manipulator for intelligent object retrieval is presented. The system was integrated using state of the art R&D hardware and software, which implemented autonomous navigation, object recognition, and object pose estimation based optimal grasping. The retrieval of an object of interest is commanded that involves subsequent object detection and recognition while autonomously navigating using the known map and starting from an arbitrary position. From close proximity, object pose estimation based optimal grasp is selected to pick up the object. The object is retrieved back to the start position in this scenario. An 84% trial-phase precision in object retrieval is achieved that can be improved using better models.
  Department: Computer Science
  Supervisor: Prof. Waqas Majeed, Dr. Arthur Choi, Dr. Sharon Perry
   |  | 
  
  

• UC-181 Prison Minecraft Game Mode Plug-In (Undergraduate Project) by , , 
  Abstract: A project designed for 最色导航's owned Minecraft server. The project centers around creating a Minecraft plug-in, a software product that is easy to activate in any Minecraft server. This plug-in changes the standard rules of Minecraft to become a classic game mode called Prison where players are taken to a special map and tasked with collecting resources in specialized mines or by fighting each other for them to earn in game currency for the purpose of buying their way to more privileged positions in the prison, gaining access to new areas and features. Prison was designed to work with KSU's current version of Minecraft and run with the Paper API used to write plug-ins like the project. In addition, the Prison plug-in features its own set of manageable plug-ins to handle separate aspects of the game mode's rules to enhance the player experience and allow for administrators to set the rules for the game and handle issues seamlessly and easily.
  Department: Computer Science
  Supervisor: Prof. Sharon Perry, Sponsors: Kylie Nowokunski, Alla Kemelmakher
   |  | 
  
  

• UC-182 Designing a User-Centered Mobile Application for Anderson Power Services (Undergraduate Project) by Guzman, Ryan, Goswick, Cooper, Phan, Anthony, Fotso, Marie, Francois, Larnel
  Abstract: This paper presents the design and development process of a mobile application for Anderson Power Services, emphasizing both frontend and backend aspects as well as their design. The frontend focuses on creating a visually appealing and user-friendly interface by utilizing clear text, an accessible color scheme, appropriate logos, animations, and modern typography. On the development side, the app leverages tools like Expo for rapid front-end development and integrates the Java-based backend with the Google Sheets API for easy data management. The backend architecture incorporates OAuth 2.0 for secure authentication, Gradle to facilitate a connection between the JavaScript frontend to the Java middleware, and Docker to connect it all. Overall, the project demonstrates the importance of balancing user-centered design principles with robust technical solutions for developing a functional and intuitive mobile application.
  Department: Information Technology
  Supervisor: Prof. Donald Privitera
   | 
  
  

• UC-184 OnAccount A Web-Based Accounting Software (Undergraduate Project) by Jackson, Manuel A, , , Powell, Zachary B
  Abstract: This project streamline and improve the efficiency of the whole accounting process, by using current best practices for user interaction engineering and current design practices. Our software should be able to provide secure, user-friendly, and accessible financial management solutions anywhere and everywhere through various devices including desktop and mobile. Allowing users to manage their accounts whenever it seems necessary while still maintaining a high level of security. The project is inspired by the various complexity and problems regarding the accounting process in the real world such as financial reporting, miscalculations, and data security; by streamlining this process and making it more automated it will mitigate the risk and problems associated with accounting.
  Department: Software Engineering and Game Development
  Supervisor: Dr. Ermias Mamo
   |  | 
  
  

• UC-186 KSU Esport: Competitive Speedrun Plugin for Minecraft Java Edition (Undergraduate Project) by , , , , 
  Abstract: The KSU Esports Minecraft Speedrun plugin transforms traditional, manually managed speedruns into an automated team-based competition event. Players are challenged to complete a set of objectives within a set time limit 鈥� promoting teamwork and strategic planning. Various modes are supported, such as weighted/unweighted speedruns, team-based speedruns, and player free-for-all. Designed for flexibility, the plugin allows for customizable settings and support for future versions of Minecraft.
  Department: Software Engineering and Game Development
  Supervisor: Supervisor: Dr. Yan Huang, Sponsor: Kylie Nowokunski; KSU Esports Team
   | 
  
  

• UC-189 ChessAI (Undergraduate Project) by , Miller, Ashton D, , Luong, Dylan, Smith, Allen L
  Abstract: Chess is a widely acclaimed two-player strategy game, where the primary objective is to checkmate the opponent's king, placing it in a position of imminent threat from which it cannot escape. Our aim was to innovate within this classic framework by developing a novel chess game that adheres to the traditional rules while enhancing accessibility for players of all skill levels. This game features a selection of AI models, each offering unique decision-making processes that create diverse gameplay experiences based on the chosen model. The AI operates by simulating every possible move on the board, meticulously evaluating each resulting position. After assessing all potential moves, it selects the optimal one based on a sophisticated or simpler algorithm/pipeline depending on the selected AI model, tailored to counter the player鈥檚 strategy. In addition to the AI-driven gameplay, we also offer a local player-versus-player mode and an online player-vs-player mode, enabling two players to compete against each other on the same computer. The ability to select and play against different AI models ensures that both casual players and serious strategists can enjoy the game in a way that suits their preferences based on their skill level.
  Department: Computer Science
  Supervisor: Prof. Sharon Perry
   |  | 
  
  

• UC-197 IT Capstone 4983: HoneyBaked Ham Intranet SharePoint Site Transformation Presentation (Undergraduate Project) by , , , , 
  Abstract: The purpose of our team鈥檚 research is to explore and define ways in which we can advance aesthetics and functionalities of how website content and ideas are presented to HoneyBaked Ham end users. Our team has goals of identifying crucial focal point areas and various ways we can overall improve upon such. We will utilize practicality, ingenuity and creativity, in order to demonstrate and perform deliveries of proper new perspectives of the site. We will seek out such advancements we can add while remaining within necessary parameters, maintaining the respected, well renowned HoneyBaked Ham Brand. We would like it to be a commonality for users to observe inventiveness, dedication and spirit, from our team refining our prototypes and the results of our academic research efforts and excursions.
  Department: Information Technology
  Supervisor: Prof. Donald Privitera
   | 
  
  

• UC-201 Halo: A Volunteer Management Application (Undergraduate Project) by , , , , 
  Abstract: Halo is a comprehensive volunteer management application (VMA) developed to address specific needs identified by Angels Among Us Pet Rescue (AAU). AAU's current system is unable to effectively manage and store complex volunteer information. As the organization鈥檚 needs evolved, AAU required a more efficient and secure solution to handle volunteer information. Our team designed Halo using the React.js framework for the frontend, the python based FastAPI framework for the backend, and a PostgreSQL database schema, with Docker containerization to ensure consistent deployment across environments. This setup enables efficient management of volunteer and team data, along with support for exporting reports in both .csv and .xlsx formats. To address security, Halo incorporates role-based access control (RBAC), differentiating access for administrators, editors, and readers. We also implemented a Google Sign-On feature with JSON web tokens (JWT) to validate users' Google login information within the database. Additionally, an Extract, Transfer, and Load (ETL) script facilitates secure data requests directly on the server. Halo鈥檚 frontend offers an intuitive interface designed to enhance user experience, with extensive search and filtering capabilities that make it easy to access and manage volunteer data. These improvements result in faster query responses and a substantial boost in usability compared to AAU's previous system.
  Department: Software Engineering and Game Development
  Supervisor: Dr. Yan Huang, Project sponsors Alla Kemelmakher and Taylor Cuffie
   |  | 
  
  

• UC-202 INDY-5 Building Map Application (Undergraduate Project) by , , Wilson, Zach W, Haas, Lucas A
  Abstract: This project鈥檚 goal is to develop a simple secure mobile application for all devices to provide a detailed interior map that can guide users to any location in the building. It will use QR codes for ease of access, and the app will provide guidance via room numbers and a visual route. Our scope includes designing the architecture of the app, creating a responsive and interactive map User interface in an app that is compatible across all devices for a nice user experience.
  Department: Computer Science
  Supervisor: Dr. Arthur Choi
   |  | 
  
  

• UC-207 Anderson Power Services Mobile Application (Undergraduate Project) by , , , , 
  Abstract: The APS Customer Experience Mobile Application streamlines and enhances customer interactions for Anderson Power Services (APS), focusing on customers who have purchased generators. This mobile application provides real-time updates, milestone tracking, and installation insights, allowing APS customers to monitor their generator installation progress with ease. By integrating Google Sheets APIs, the app enables seamless data synchronization, providing accurate and timely updates on generator status. With its cross-platform support on iOS and Android, the application reduces the need for manual communication, improving customer satisfaction and operational efficiency by 20%.
  Department: Software Engineering and Game Development
  Supervisor: Dr. Yan Huang
   |  | 
  
  

• UC-221 Accounting Treasury Industries Web Application (Undergraduate Project) by , , ,
  Abstract: This accounting software project is designed to provide a comprehensive and efficient solution for financial management within an organization. By focusing on ease of usability, accuracy, and compliance, the software enables users to record, manage, and analyze accounts, journals, and financial transactions seamlessly. Core features include transaction journalization, chart of accounts setup, financial statement generation, and robust account management, all of which are supported by strong data validation and secure access controls. This system seeks to streamline accounting workflows, minimize human/manual errors, and enhance user experience. The ultimate aim is to deliver an intuitive yet powerful tool that supports effective financial management and strengthens operational efficiency.
  Department: Software Engineering and Game Development
  Supervisor: Dr. Ermias Mamo
   |  | 
  
  

• UC-223 Cybersecurity Website Hardening Project (Undergraduate Project) by , , vandorn, elijah, , 
  Abstract: The project aims to secure the Akwaaba website using Apache, MariaDB, Red Hat OS, and PHP on a virtual machine. It includes assessing assets and vulnerabilities, implementing security policies, and conducting a red/blue team exercise for ethical hacking experience.
  Department: Information Technology
  Supervisor: Prof. Donald Privitera
   | 
  
  

• UC-225 Golf Course Pace Management Simulation (Undergraduate Project) by Miller, Ashton D
  Abstract: The game of golf has been played for centuries so it has seen handfuls of evolutions throughout its time being played. Throughout the games evolution one factor of its existence has hardly ever changed, time. In current day golf standards, time and the management of time is a significant part of not only how well the game as a whole run but how the courses that own venues to play the game operate as well. In golf there is a standard for time known as the pace of play model, where groups that are sent off during the day are queued into a course by whatever hole they're told to start on. If the golfers fall behind the pace of play standard, then largely, the courses flow of players that bring them revenue decreases in pace. Once a course experiences this decrease in pace, a large hit is taken to the course's financials. This widely affects everyone involved in the course's operation from cart attendants to restaurant workers, and lastly but not at all in the least, the management. These affects are not only present in a normal day of golf but also in tournament settings as well where groups finishing later in the tournament time frame may cost cart expenditures for the rest of the course not playing in a tournament and wanting to go out and get in a normal round. The project shown will display a knowledge of how this model takes precedence in the world of golf by displaying groups along with their carts queueing into the course structure and playing through a tournament. The goal of the project is to show the user the frequency of pace violations that occur and marshal interference needed in order to keep up the pace of the tournament so that the course can study time discrepancies and find where on the course these discrepancies occur and with what groups they occur in. In golf it is imperative that the pace of play model be successfully followed, otherwise structurally related to time, the course takes a hit both organizationally and financially.
  Department: Computer Science
  Supervisor: Prof. Christopher Regan
   | 
  
  

• UC-226 Real-Time Bus Monitoring Using Kafka (Undergraduate Project) by , , , Pruitt, Brian A, 
  Abstract: The GCPS Real-Time Bus Monitoring System aims to enhance bus operations for Gwinnett County Public Schools by transitioning from a polling-based system to a real-time Kafka event-streaming architecture. This project processes telemetry data from over 2,000 buses, simulating a scalable, near-instantaneous data flow into an SQL Server database. Key features include real-time data validation, efficient data storage, and containerized deployment for consistency across environments. Using an Agile approach, our team handled evolving requirements from the sponsor, who is new to senior project collaborations. This system enables GCPS to monitor bus locations with reduced latency, enhanced accuracy, and improved resource management, laying a robust foundation for future scalability and analytics.
  Department: Computer Science
  Supervisor: Prof. Sharon Perry
   |  | 
  
  

• UC-231 Symptom-Based Disease Prediction (Undergraduate Project) by Barber, Jarred M, Clark, Kody, Mwangi, Ryan, Zheng, Zhiwen
  Abstract: This project focuses on leveraging large-scale data sets and advanced analytical techniques to predict the onset of diseases. By integrating data from medical records, genetic information, and environmental factors, the project aims to identify patterns and risk factors associated with various diseases. Machine learning algorithms and statistical models are employed to enhance the accuracy of predictions, enabling early diagnosis and personalized healthcare interventions. This approach improves patient outcomes and contributes to the efficiency and effectiveness of healthcare systems.
  Department: Computer Science
  Supervisor: Dr. Dan Lo
   | 
  
  

• UC-242 AC-10 AI & Music Processing (Undergraduate Project) by , , Wilson, Sterling J, Egwuatu, Michael
  Abstract: There is a broad range of styles and philosophies, for teaching young children how to play music. Some are based on repetition and memorization of songs, and others build up a foundation of musical patterns and motifs. Arguably, the latter style, will better develop the skills needed for improvisation and composition of new music. Inspired by this observation, we aim to improve the ability of (recurrent) neural networks to synthesize music based on a more careful training.
  Department: Computer Science
  Supervisor: Dr. Arthur Choi
   |  | 
  
  

• UC-247 Using Dynamic Difficulty Adjustment (DDA) to improve health and wellness apps and programs (Undergraduate Project) by 
  Abstract: Physical inactivity, obesity and Type 2 Diabetes cost the United States鈥� economy more than $700 billion a year (CDC). Yet, individuals spend $137 billion dollars a year on gym memberships to get in shape and feel better, without attaining results and dropping out. 鈥溾��63% of new members will abandon activities before the third month, and less than 4% will remain for more than 12 months of continuous activity.鈥� (Sperandei et al). The personal training apps don鈥檛 fare better, with 71% of users disengaging within 90 days (Amagai et al). The higher chances of people dropping out are due to "a higher degree of discomfort and distress during exercise sessions" (Sperandei 919). Additionally, individuals with less than 2 training sessions per week have higher attrition rates (Garay et al 7). Our hypothesis is that Digital Difficulty Adjustment (DDA) could be used beyond videogames to create positive habits and to increase the amount of physical exercise by making the exercises鈥� intensity levels adapt to the physical levels of the person exercising in real-time. DDA is a technique used in video games to adaptively change the game's difficulty level in response to the player's performance and creates an engaging and tailored playing experience that lasts longer for the player. We expect the findings of this research can be applied to designs in other areas of healthcare and wellness programs to effectively improve adherence, and reduce attrition of these programs, potentially reducing the national and personal costs in poorly designed digital health and wellness products.
  Department: Software Engineering and Game Development
  Supervisor: Dr. Lei Zhang
   | 
  
  

• UC-248 Campus AI Companion Mobile App (Undergraduate Project) by , , , , 
  Abstract: The Campus AI Companion app is designed to enhance students' university experiences by providing personalized recommendations for courses, events, clubs, and career paths. Leveraging OpenAI鈥檚 language model and developed using React Native, this mobile application integrates academic and social guidance, tailored for individual users based on their interests and performance. This AI-driven companion aims to help students better navigate their university journey by providing seamless access to resources, activities, and support that align with their academic and personal goals.
  Department: Computer Science
  Supervisor: Prof. Sharon Perry
   |  | 

Academic courses graduate (e.g. capstones, games, innovative special topics projects, other course projects)

• *GMC-130 School Bus Monitoring Simulation w/ Apache Kafka (Graduate Project) by , , , , Lois, Julio
  Abstract: Our project is a proof-of-concept of event-streaming bus telemetry data using Apache Kafka (Kafka), as requested by our client, Gwinnett County Public Schools (GCPS). The Kafka event stream is more efficient than GCPS鈥檚 current process of pulling bus telemetry data: calling APIs every 5 seconds. Moving to Kafka will provide GCPS near real-time insights into bus locations and speeds, giving visibility to whether buses drive safely and punctually. Our simulation produces synthetic bus data to be passed through Kafka and consumed. It features two UIs for system monitoring and data visualization.
  Department: Software Engineering and Game Development
  Supervisor: Dr. Reza Parizi
   | 
  
  

• GMC-137 IoT Security Vulnerabilities and How to Improve Them (Graduate Project) by , , , 
  Abstract: With the increased usage of IoT devices in homes as well as different industries, vulnerabilities have also increased significantly. The IoT devices are small in size, and it is hard to incorporate security in the software because security has high demand for computation. We have been conducting this research in order to find more suitable security methods that are lightweight as well as efficient. We have decided to move away from key hiding algorithms, which have increased time and space consumption, in favor of smaller and quicker block cipher algorithms.
  Department: Computer Science
  Supervisor: Dr. Dan Lo
   | 

• GMC-146 Legal-Insight - Legal text summarizer (Graduate Project) by , , 
  Abstract: Many people struggle to fully understand complex legal documents, such as terms of service agreements, contracts, and privacy policies, due to dense jargon, small fonts, and lengthy paragraphs that make critical information difficult to grasp. This lack of clarity can lead individuals to inadvertently agree to terms they might not fully understand or miss important clauses. Recognizing these challenges, we developed LegalInsight to make legal information more accessible and comprehensible. LegalInsight simplifies lengthy legal documents by creating clear and concise summaries, allowing users to easily digest essential information. It also includes an interactive Q&A feature where users can ask specific questions about a document or its summary, receiving targeted responses that clarify confusing sections or terms. This tool serves a wide range of users鈥攆rom students and elderly individuals to large corporations鈥攂y providing a user-friendly interface that makes navigating complex legal texts easier.
  Department: Computer Science
  Supervisor: Dr. Md Abdullah Al Hafiz Khan
   |  | 
  
  

• GMC-157 Text-to-Digital Person Video Generator: DigitalAvatarGen (Graduate Project) by , , , , 
  Abstract: The Text-to-digital person video generator: DigitalAvatarGen project uses AI to create lifelike videos of 2D digital avatars from user text input. Users enter text, select a voice and select or upload an avatar, and generate a video using DigitalAvatarGen web application which uses Google TTS and SadTalker, to synchronize voice, expressions, and lip movements. Key contributions include a customizable user interface, personalized voice and avatar options, and an optimized backend for efficient video generation. This tool provides an engaging, realistic solution for applications in education, media, and customer interaction.
  Department: Information Technology
  Supervisor: Dr. Ying Xie
   |  | 
  
  

• GMC-158 Evaluating TCP Protocol Performance in Cloud Environments (Graduate Project) by 
  Abstract: This research investigates the performance of four different TCP algorithms鈥擝BR, Reno, Vegas, and Cubic in a high-latency and congested condition within a cloud-based environment using EC2 instances and Mininet for network simulation. The study aims to evaluate the throughput and congestion window (cwnd) behavior of each algorithm under various network conditions to identify their strengths and weaknesses. By analyzing the performance metrics across different TCP algorithms, we provide insights into their suitability for cloud infrastructure, contributing to optimized network protocol choices for cloud-based applications and services. The results offer valuable guidance for enhancing network performance in dynamic cloud environments.
  Department: Computer Science
  Supervisor: Dr. Ahyoung Lee
   | 
  
  

• GMC-168 Hybrid Approach of Data Mining and Deep Learning for Network Intrusion Classification in Big Data (Graduate Project) by , , Gurram, Yaswanth srinivas
  Abstract: The growing complexity and volume of network traffic pose significant challenges to traditional intrusion detection systems (IDS), often leading to inefficiencies in detecting unauthorized access and malicious activities. To identify different types of network attacks, many intrusion detection systems (IDSs) have been proposed using artificial intelligence or machine learning, but the results are still not satisfactory for most of these systems. Recently in some research, deep learning models have shown promising performance in big data analysis. However, a combined data mining and deep learning approach in big data for the detection of intrusions has not been scrutinized. This research aims to develop a hybrid approach combining data mining techniques and deep learning models to improve the detection of intrusions in large-scale networks. In this paper, we propose a Genetic Algorithm (GA) and Minimum Redundancy Maximum Relevance (mRMR) to select optimal features by reducing the dimensionality of the dataset. Initially (mRMR) selects features based on high relevance to the target variable and also has the minimum overlapping information of those selected features. Then GA algorithm finds the best subset of features where it evaluates the various combinations of attributes and chooses the best ones to enhance the performance of the model. After that, the deep learning model Convolutional Neural Network (CNN) was introduced, which uses 1D convolutional layers to detect small, localized complex patterns by adopting structured data. By leveraging data mining for feature extraction and deep learning for anomaly detection, the proposed system seeks to enhance the accuracy and efficiency of IDS in handling big data. The expected results include improved detection rates, reduced false positives, and robust performance in processing large network intrusion datasets.
  Department: Computer Science
  Supervisor: Dr. Chen Zhao
   | 
  
  

• GMC-191 Optimizing K-means Clustering for Customer Analytics: A Multi-faceted Enhancement Approach (Graduate Project) by , , Bonigala, Dedeepya
  Abstract: This paper presents a detailed analysis of the al- gorithmic complexity of the K-means clustering algorithm, a foundational method in unsupervised machine learning. Although the problem of finding the optimal solution is NP-hard, K-means is widely used for efficiently partitioning data into clusters by minimizing within-cluster variance. We explore four main ideas for improvement:1) parallel points generation and processing for speeding up convergence, 2) penalty scoring for avoiding clusters with high variability within them, 3) utilization of other distance measurements such as Manhattan distance for providing better clustering in structures of objects of different nature and 4) probability addition in the form of Gaussian Mixture Models (GMM) for more adaptable and soft k-means clustering. A practical application of K-means is applied to telecommu- nication data to understand purchasing behavior via customer segmentation. It was observed that K-means++ provided the most optimum method for centroid initialization while parallel K-means performed the task of minimizing the execution time. Penalty scoring produced more balanced clusters compared with the baseline and GMM allowed for more flexibility in defining cluster boundaries. Initial findings show that K-means++ has an average silhouette score of 0.67 while the GMM method has an average silhouette of 0.65 which is particularly more appropriate for the complex customer segmentation.
  Department: Computer Science
  Supervisor: Dr. Chen Zhao
   | 
  
  

• GMC-2162 Prompt Engineering and its Effects On AI and Human Relationships: A Contemporary Approach (Graduate Project) by , , Thallapally, Nivesh
  Abstract: A. Background: Prompt engineering refers to the process of designing and refining input prompts for AI models (especially language models like GPT) to improve their outputs. It has become a critical tool in maximizing the performance and utility of AI models in diverse applications, from customer service to content creation. Beyond technical aspects, the interaction between humans and AI is increasingly shaped by the effectiveness of these prompts. B. Motivation: As AI becomes more integrated into daily life, the way humans interact with AI models is profoundly influenced by prompt engineering. Misaligned prompts can lead to misunderstanding, confusion, or unintended outcomes, affecting both the utility of AI systems and the trust people place in them. Our project seeks to understand how different prompt strategies impact not only AI performance but also human perceptions and relationships with AI systems. By exploring these dynamics, we aim to develop best practices in prompt engineering that foster both efficient AI performance and positive human-AI relationships. C. Expected Results: We expect to demonstrate that well-constructed prompts not only improve AI output quality but also lead to more transparent, trustworthy, and meaningful human-AI interactions. This will be quantified through various metrics such as response accuracy, user satisfaction, and interaction smoothness.
  Department: Computer Science
  Supervisor: Dr. Chen Zhao
   | 
  
  

• GMC-218 PET RESCUE AI-BASED SUPPORT APPLICATION (Graduate Project) by , , , , 
  Abstract: This project focuses on the development of an AI-driven support application for foster caregivers at Angels Among Us Pet Rescue. The application provides foster caregivers with real time assistance through an interactive chatbot, task reminders and resource management capabilities, streamlining the caregiving process. By leveraging automation and AI, the application enhances both the foster experience and operational efficiency aligning with the organizations mission of improving animal care.
  Department: Information Technology
  Supervisor: Dr. Jack Zheng, Project Owner: Angels Among Us Pet Rescue
   |  | 
  
  

• GMC-219 Athlete-Agent Connect Mobile App (Graduate Project) by , , 
  Abstract: The Athlete-Agent Connect app aims to bridge the gap between athletes and agents, simplifying the process of professional engagement. By providing a digital space for talent acquisition and event coordination, the app fosters networking, recruitment, and collaboration within the sports industry. The platform鈥檚 features are tailored to meet the needs of athletes looking for representation and agents seeking clients, with tools for direct communication, event planning, and a calendar of relevant sports gatherings. This mobile app serves as a dedicated platform for athletes and sports agents to connect, collaborate, and enhance professional opportunities. The app enables athletes to hire agents for representation, allows agents to offer their services, and supports the creation of sporting events where athletes and agents can meet in person. Additionally, users can browse and RSVP to upcoming events related to their sports network.
  Department: Information Technology
  Supervisor: Dr. Ying Xie
   |  | 
  
  

• GMC-241 GTA Request / Hiring Management (Graduate Project) by , , , , 
  Abstract: The Graduate Teaching Assistant (GTA) Management System is designed to address inefficiencies in the GTA hiring and assignment process at the departmental level. This full-stack web application utilizes modern Python frameworks such as Flask for backend development, providing a robust and scalable foundation for data handling and business logic. The frontend is developed using HTML, CSS, and JavaScript, ensuring an intuitive and responsive user experience. Key features include user access, automated GTA request generation based on course catalog data, and real-time tracking of hiring progress. By integrating data import capabilities from sources like Owl Express via Excel, the system handles complex requirements like cross-listed sections and enrollment rules. This application aims to enhance administrative efficiency, reduce workload for faculty and staff, and improve the coordination and transparency of GTA management processes across departments. Through leveraging the power of Python frameworks and full-stack development methodologies, the system is both flexible and scalable, ready to adapt to future institutional needs.
  Department: Information Technology
  Supervisor: Course Instructor- Dr. Jack Zheng, Project Sponsor- Dr. Zhigang Li
   |  | 
  
  

• GMC-246 Enhancing Workforce Management through Advanced HR Analytics (Graduate Project) by , Gurram, Ruthvik Reddy
  Abstract: The business analytics of employee data is a concern that human resource departments worldwide deal with. Some big organizations have entire teams working on analyzing these metrics. To obtain insights about employee turnover rates, performance trends, and compensation patterns from data, the Data warehousing techniques鈥擮LAP and ETL鈥攃an be used to handle data. This paper aims to develop an OLAP model for multi-dimensional analysis using data warehousing techniques that help extract valuable insights from the data. Popular datasets will be used, and the model will be evaluated according to standards.
  Department: Computer Science
  Supervisor: Dr. Chen Zhao
   | 
  
  

• GMC-4190 CellNucleiRAG - Smart Search Tool for Cell Nuclei Research (Graduate Project) by 
  Abstract: CellNucleiRAG is a specialized tool developed to address a significant challenge in medical research: the rapid retrieval and synthesis of detailed information on cell nuclei. Understanding cell nuclei characteristics is crucial in fields like pathology, oncology, and diagnostics, where detailed cell analysis can guide disease identification and treatment planning. However, accessing relevant, organized information on specific cell nuclei types, datasets, models, and methods is often time-consuming, requiring manual searches through multiple, disparate sources. CellNucleiRAG solves this problem by acting as a smart search engine, designed specifically for cell nuclei research, combining traditional retrieval methods with advanced AI capabilities. Built with an underlying Retrieval-Augmented Generation (RAG) architecture, CellNucleiRAG leverages MinSearch for rapid data retrieval, pulling relevant records from a curated dataset that contains information on various nuclei types, datasets, and analytical models. Once relevant data is retrieved, it is processed by an LLM (Large Language Model) to generate contextually accurate, human-readable responses. This dual approach ensures both precision and clarity, allowing researchers to receive comprehensive answers rather than isolated data points. Key technologies used in this project include Docker, for environment consistency; Flask, for a streamlined user interface; PostgreSQL, for storing interactions and user feedback; and Grafana, for real-time system performance monitoring. User feedback is incorporated to continually refine the tool, enhancing the accuracy and relevance of responses.
  Department: Computer Science
  Supervisor: Dr. Coskun Cetinkaya
   | 

Research projects completed by undergraduate students.

• UR-147 An 8-bit Digital Computer Design & Implementation (Team COA-WM1) (Undergraduate Research) by Sherard, Adrian L, Flores, Jesus, Lamsal, Biswash, Hammontree, Blake, Pitts, William
  Abstract: 8 bit computer design using NI multisim
  Department: Computer Science
  Supervisor: Prof. Waqas Majeed
   | 
  
  

• UR-172 A Comparative Study of LLM Effectiveness in Mental Health Assistance (Undergraduate Research) by 
  Abstract: This study evaluates the effectiveness of LLMs in supporting mental health applications by analyzing their performance in understanding and categorizing user (mental health-related) inputs. We collected data from various mental health apps on the Google Play Store, including user reviews and app descriptions, and filtered content using a targeted mental health keyword bank. Sentiment analysis and keyword similarity scores were generated for reviews using RoBERTa-based models, this showed us how each review aligned with the mental health keywords advertised by the app and how users felt about the app. We prompted four modern LLMs: GPT-4o, Claude 3.5 Sonnet, Gemma 2, and GPT-3.5-Turbo. We provided Gemma 2 and GPT-3.5-Turbo with our dataset for more informed outputs. Our prompts consisted of five common mental health conditions (depression, anxiety, ADHD, PTSD, and insomnia) and we asked for the models to provide us with up to five app recommendations. The results showed that our data-enhanced LLMs noticeably outperformed the other state-of-the-art LLMs in accuracy, quality, and variety of outputs while being much more cost-effective. This suggests that data-enhanced, low-cost LLMs can serve as an effective alternative to newer, more powerful, and more expensive models, achieving notably better results in interpreting nuanced text for mental health applications.
  Department: Computer Science
  Supervisor: Dr. Md Abdullah Al Hafiz Khan
   | 
  
  

• UR-213 Generative AI & Cybersecurity (Undergraduate Research) by , ,
  Abstract: This research project details the impact of Generative AI on Cybersecurity through both its potential enhancements and threats. Using advanced AI algorithms, this project explores how Generative AI can strengthen cybersecurity through systems like Anomaly Detection, Intrusion Detection Systems (IDS), and Malware Analysis. Also, this project addresses the growing challenges posed from Generative AI. In particular, issues surrounding Deepfake Phishing and Polymorphic Malware are discussed. Solutions to mitigate these issues are also provided to engage further understanding in the field. The goal of this research is to offer practical solutions for addressing the growing field of AI-driven cybersecurity.
  Department: Computer Science
  Supervisor: Dr. Yong Shi, Project Advisor: Prof. Sharon Perry
   |  | 
  
  

• UR-239 Human-AI Annotator Tool (Undergraduate Research) by , , ,
  Abstract: The HK-01 Human-AI Annotator Tool is a web-based system developed to facilitate the annotation of Electronic Health Records (EHRs) for mental and behavioral health, using ICD-10 codes. The tool allows multiple expert annotators to tag critical information, making it easier to catalog patient data accurately. Future plans include integrating AI to streamline and scale the annotation process, improving both efficiency and accuracy.
  Department: Computer Science
  Supervisor: Dr. Md Abdullah Al Hafiz Khan, Dr. Arthur Choi, Prof. Sharon Perry
   | 

Research projects completed by master's students.

• GMR-124 Emotion-Based Synthetic Feature Binary Classification of Human vs LLM Generated Text/Essay (Master's Research) by , , ,
  Abstract: Sentimental analysis is a popular method to classify text into various emotional tones and intentions. In the meanwhile, the emergence of large language models (LLMs) has become ever more capable, their potential to cause harm through information fabrication, misleading propagation, or mere lack of capability has also increased. Therefore, our project is designed to discover any patterns that could potentially uncover texts origins of human and LLM during sentimental analysis. Our dataset covers over 57,000 lengthier essay samples (70% human vs 30% LLM), we use the state-of-art pre-trained DistilRoBERTa-base, a powerful pre-trained language model that is a more condensed and speedier version of Google's BERT, as our bidirectional transformer. Moreover, we plot the experiment results in histograms and statistical analysis and propose potential future research directions.
  Department: Computer Science
  Supervisor: Dr. Md Abdullah Al Hafiz Khan
   | 
  
  

• GMR-159 LLM enabled Synthetic dataset generation for Human-AI teaming Algorithm (Master's Research) by Potluri, Sai Sanjay
  Abstract: This research explores using Large Language Models (LLMs) to generate synthetic datasets for Human-AI teaming algorithms, focusing on mental health assessments. We create a diverse dataset simulating human-AI collaboration scenarios in diagnostic processes. The synthetic data is labeled through an innovative approach involving two human annotators and three LLMs, using majority voting for consensus-based annotations. This dataset serves as a resource for training and evaluating Human- AI teaming algorithms, enabling exploration of collaboration dynamics between human expertise and AI in complex decision-making. Our approach addresses the scarcity of real-world data in Human-AI teaming scenarios and provides a controlled environment for algorithm development, potentially accelerating advancements in this field.
  Department: Computer Science
  Supervisor: Dr. Md Abdullah Al Hafiz Khan
   | 
  
  

• GMR-165 An Empirical Study of Prompt-based Non-functional Requirements Classification (Master's Research) by Kim, Allen
  Abstract: In modern software development, Non-Functional Requirements (NFR) are essential to satisfy users鈥� needs. Distinguishing different categories of NFR is tedious, error-prone, and time consuming due to the complexity of software systems. In our project, we conducted a comprehensive study to evaluate the performance of prompt-based NFR classification by designing various handcraft templates and soft templates on the pre-trained language model (i.e., BERT). Our experimental results show that handcraft templates can achieve best effectiveness (e.g., 83.52% in terms of F1 score) but with unstable performance for different templates.
  Department: Software Engineering and Game Development
  Supervisor: Dr. Xia Li
   | 
  
  

• GMR-196 Integrated Sentiment and Behavioral Analysis of Online Product Reviews (Master's Research) by , 
  Abstract: The "Integrated Sentiment and Behavioral Analysis of Online Product Reviews" project helps businesses gain actionable insights from Product reviews by combining sentiment and behavioral analysis using NLP models like VADER and BERT. This dual approach categorizes reviews as positive, neutral, or negative and identifies themes such as preferences and complaints through Named Entity Recognition and topic modeling. By capturing both the emotional tone and specific product feedback, this method highlights consumer likes and pain points, assisting in targeted improvements for product design and customer service. The project addresses challenges in analyzing complex expressions like sarcasm, providing a robust framework for extracting meaningful insights from vast amounts of review data. Adaptable across various datasets, the model offers scalable benefits for enhancing e-commerce strategies through data-driven decisions based on real consumer feedback.
  Department: Computer Science
  Supervisor: Dr. Md Abdullah Al Hafiz Khan
   | 
  
  

• GMR-208 Automatic Categorization of Behavioral Health Issues in Police reports (Master's Research) by , , 
  Abstract: 911 is often the first place contacted for dealing with behavioral health related (BHR) issues. Its estimated at least a fifth of all calls are related to behavioral health, and with BHR affected convicts having a recidivism rate of around 30%, its not hard to see how straining these issues can become on systems already stretched thin, where chronic understaffing is often a reality. A great solution would be if we could intervene as soon as possible to get people the treatment they need, police reports would be excellent for identifying and treating these individuals, but annotation is a long tedious task only certain people have security clearance to do and as mentioned earlier departments are often understaffed. That is why with the help of keywords given to us by behavioral health professionals, we have developed a model for automatic categorization of police reports that can classify police reports into several categories of class type (Situation, Situation Mental Health, Child, Disposition, Disposition Mental Health, Drugs, Medication, Medication Mental Health) by learning the correlation between co-occurrences of class types given keywords, evidence type given keywords, and class type given keywords and then combining those with the embeddings of a Feed Forward Network that analyzed relevant sentences from reports. With this model we were able to achieve an accuracy rate of 72% which was significantly higher than other state of the art methods typically used.
  Department: Computer Science
  Supervisor: Dr. Md Abdullah Al Hafiz Khan, Dr. Yong Pei
   | 
  
  

• GMR-210 Cogni-Resource: AI-Driven Reflective Feedback Analysis for Enhanced Learning Insights and Resource Discovery (Master's Research) by 
  Abstract: Cogni-Resource is a unified platform enhanced by AI that merges the introspective analysis of Cogni-Reflect with the precise resource exploration functions of the Learning Resource Finder, providing a holistic tool to improve educational environments. The Cogni-Reflect component uses advanced Large Language Models (LLMs) to examine student reflections, giving educators instant insights into learning results, difficulties, and areas where students may require extra assistance. Cogni-Reflect allows instructors to adjust their teaching by analyzing key themes and topics in reflective narratives, leading to a more adaptive and successful learning atmosphere. Using both web scraping and OpenAI API integration, the Learning Resource Finder gathers educational content tailored to the specific needs of students. This tool sifts through and pairs up useful resources, like articles, tutorials, and research papers, with the exact topics and learning goals students are focusing on, getting rid of unimportant content and offering fast access to top-notch materials. Collectively, these instruments make up Cogni-Resource, a platform that not only simplifies the typically lengthy process of reflection analysis for teachers but also enables students to autonomously delve into selected, tailored materials. Cogni-Resource promotes an educational setting where both assessing learning advancement and finding personalized materials are made easier, emphasizing both instructional excellence and student independence. Educational institutions that utilize Cogni-Resource have the ability to take a comprehensive approach to analytics and content delivery. This allows educators to make informed, timely modifications to their teaching methods while also assisting students with their self-directed learning paths. In the end, Cogni-Resource connects reflective analysis and resource accessibility, improving educator interventions and student learning outcomes with the help of AI.
  Department: Computer Science
  Supervisor: Dr. Nasrin Dehbozorgi
   | 
  
  

• GMR-215 Efficient Sentiment Analysis using Encoder-only Transformer (Master's Research) by , , 
  Abstract: In the era of social media, sentiment analysis has emerged as a vital instrument for comprehending public opinion, especially on sites like LinkedIn and Twitter. Because user-generated content is informal and noisy, traditional sentiment classification techniques like Naive Bayes and Support Vector Machines sometimes find it difficult to capture context, sarcasm, and long-term interdependence. In order to improve sentiment analysis accuracy for social media datasets with a specific focus on sentiments related to corporate layoffs, this study suggests an encoder-only transformer model. Our method successfully captures intricate phrase patterns and contextual subtleties in textual data by leveraging the self-attention mechanism built into transformer designs. To assess the model鈥檚 performance on unseen data, we used the LinkedIn dataset for testing and the Twitter dataset for training. To help the model understand the semantic linkages in the text, we used Word2Vec for tokenization and representation. Our research indicates that the transformer model works noticeably better than conventional sentiment analysis methods, exhibiting enhanced resilience and flexibility when dealing with colloquial language and conflicting emotions. This development could have an impact on businesses and organizations looking to use social media insights on layoffs to make data-driven decisions.
  Department: Computer Science
  Supervisor: Dr. Chen Zhao
   | 
  
  

• GMR-216 AI/ML-Based Water Quality Monitoring Mobile App for Predicting E.coli in Surface Waters (Master's Research) by 
  Abstract: E.coli contamination in surface waters has proven to be a significant public health concern, requiring innovative monitoring solutions. This paper presents the design of an AI-driven mobile application to predict whether E.coli bacteria are present at levels exceeding acceptable thresholds in surface waters. The methodology employs sensor devices to collect water quality data parameters, such as water temperature, pH, dissolved oxygen, and turbidity. A dataset is generated based on these parameters, and machine learning (ML) algorithms are applied to evaluate accuracy, precision, recall, and processing time. Additionally, our ML algorithms establish a correlation matrix among water quality parameters to identify the key factors influencing E.coli levels. We applied various machine learning techniques to the dataset, including Support Vector Regression (SVR), Random Forest Classification (RFC), XGBoost, and ensemble methods that combine these algorithms. Our findings indicate that the ensemble of Random Forest Classification and XGBoost achieved the highest accuracy. Users can view E. coli predictions based on current sensor values through our Mobile App.
  Department: Software Engineering and Game Development
  Supervisor: Dr. Ahyoung Lee
   | 
  
  

• GMR-229 Semantic Search using Sentence Transformers (Master's Research) by , 
  Abstract: Traditional keyword-based search engines struggle to accurately capture the semantics of user queries in today's enormous digital resources. Our research study focuses on creating a semantic search engine that uses Sentence Transformers to improve information retrieval by understanding the context of queries and documents. Our method creates sentence embeddings for documents and user queries, allowing retrieval based on semantic similarity rather than keyword matching. The project involves data collection and preprocessing, feature extraction with Sentence Transformers, and implementation of a search engine that ranks documents based on cosine similarity to query embeddings. According to preliminary testing, this method greatly improves search relevancy and accuracy, we have compared the results with a baseline algorithm, BM25, to assess the effectiveness of Sentence Transformers in enhancing retrieval relevance. This work opens the door for future refinements in retrieval systems based on natural language processing and shows how semantic search engines can deliver results that are more contextually aligned.
  Department: Computer Science
  Supervisor: Dr. Md Abdullah Al Hafiz Khan
   | 
  
  

• GMR-4234 Evaluating Instance Segmentation Models on Histopathology Datasets (Master's Research) by 
  Abstract: Instance segmentation is transforming digital pathology by enhancing the speed and accuracy of tissue sample analysis through advanced image processing techniques. Whole Slide Imaging (WSI) converts traditional microscope slides into high-resolution digital formats, enabling detailed examinations. This paper presents a brief experimental survey of instance segmentation models on two prominent histopathology datasets: PanNuke and NuCLS. Unlike previous surveys that merely describe deep learning models for general pathology images, we conduct experiments using state-of-the-art models including Mask R-CNN, Detectron2, YOLOv8, YOLOv9, and HoverNet on both datasets. Our study evaluates these models for both binary and multiclass instance segmentation tasks. The NuCLS dataset, featuring over 220,000 annotated nuclei from breast cancer histopathology images, is used for multiclass segmentation across 13 distinct nuclear classes. The PanNuke dataset, comprising 205,343 labeled nuclei across 19 tissue types, is employed for both multiclass and binary instance segmentation of five cell types: neoplastic, inflammatory, soft tissue, dead, and epithelial. We assess each model's performance using metrics such as mean average precision (mAP), F1 score, and Dice coefficient, providing a comprehensive evaluation of their strengths and limitations. The results of our study offer valuable insights into the capabilities of different instance segmentation models in histopathology image analysis. We observe varying performance across tissue types and cell categories, highlighting the importance of model selection based on specific histopathology tasks. Our findings aim to guide researchers in choosing appropriate models for their specific needs, ultimately contributing to the advancement of digital pathology and improving diagnostic accuracy in clinical practice. Also provides a foundation for future research in instance segmentation for histopathology images.
  Department: Computer Science
  Supervisor: Dr. Sanghoon Lee
   | 
  
  

• GMR-7175 Enhancing Alzheimer鈥檚 Diagnosis through Spontaneous Speech Recognition: A Deep Learning Approach with Data Augmentation (Master's Research) by Mutala, Venkata Sai Bhargav
  Abstract: Alzheimer鈥檚 disease (AD) is a growing public health issue due to its progressive nature and rising prevalence. This study explores a neural network model trained on speech data from the ADReSS2020 Challenge dataset to distinguish AD patients from healthy individuals, using log-Mel spectrogram features. To improve accuracy, five data augmentation methods, including pitch and time shifting, were used. The results highlight deep learning, combined with data augumentation, as a promising, scalable, and noninvasive approach for early AD diagnosis
  Department: Information Technology
  Supervisor: Dr. Seyedamin Pouriyeh
   | 
  
  

• GMR-7179 Improving Alzheimer鈥檚 Detection via Synthetic Data Generation Using GPT-4 and Multi-Level Embeddings (Master's Research) by Mutala, Venkata Sai Bhargav, Shahid, Imaan
  Abstract: This study leverages large language models (LLMs), particularly GPT-4, to overcome the data limitations often encountered in Alzheimer鈥檚 detection. We utilize GPT-4 for data augmentation, generating synthetic speech transcripts to enhance machine learning model training. Our approach combines fine-tuned BERT embeddings with CLAN-derived linguistic features, as well as sentence-level embeddings, to improve classification performance on the ADReSS2020 dataset. BERT and CLAN features capture detailed linguistic variants, while sentence embeddings offer robust semantic representations, collectively enhancing the accuracy and generalization of the models. Among the classifiers tested, the Random Forest model shows the best performance, achieving an accuracy of 88% with sentence embeddings, surpassing other models in detecting Alzheimer鈥檚 from speech patterns. The integration of LLM-augmented data and multilevel embeddings presents a promising solution to the data scarcity issue in medical research, enabling more accurate and reliable Alzheimer鈥檚 diagnoses.
  Department: Information Technology
  Supervisor: Dr. Seyedamin Pouriyeh
   | 
  
  

• GMR-8193 Harnessing ML-Powered HPCC Systems for Advanced Cybersecurity Analytics (Master's Research) by 
  Abstract: Information security in the era of AI and automation is the biggest challenge for cybersecurity professionals. Traditional information security protection has limitations in detecting zero-day attacks, which can be overcome with machine learning-based information security. An ML-powered intrusion detection system uses statistical analysis to spot deviations from normal behavior and helps to detect new and unknown threats. This poster will demonstrate how an open-source platform can be used for cybersecurity by leveraging various machine-learning algorithms.
  Department: Software Engineering and Game Development
  Supervisor: Dr. Seyedamin Pouriyeh
   |  | 
  
  

• GMR-8195 Machine Learning-Enhanced HPCC Systems for Alzheimer's Disease Detection: A Scalable Solution for Early Diagnosis (Master's Research) by 
  Abstract: Alzheimer's disease is an incurable brain disorder that gradually deteriorates memory and cognitive abilities, leading to symptoms such as memory loss, confusion, difficulty in thinking, and changes in language, behavior, and personality. Early diagnosis that is effective, innovative, and cost-efficient can help mitigate damage to nerve cells. Detecting these symptoms through voice responses and analyzing the corresponding transcripts offers a promising approach. This poster demonstrates how an open-source platform can be utilized for text classification to identify Alzheimer's disease, employing fast, inexpensive, and non-invasive methods to complement other diagnostic techniques
  Department: Software Engineering and Game Development
  Supervisor: Dr. Seyedamin Pouriyeh
   |  | 

Research projects completed by PhD students. 

• GPR-1194 Computer Vision-Enhanced Spectroscopy for Glucose Prediction: An In Vitro Validation Study (PhD Research) by 
  Abstract: This study introduces a novel computer vision-based spectral approach for non-invasive glucose detection using synthetic blood samples. We developed an experimental setup with glucose concentrations from 70 to 120 mg/dL, using two dye methods. Light sources tested included an 850 nm LED, 850 nm laser, 808 nm laser, and 650 nm laser, with image capture via a 1080p IR camera. Data augmentation, including Gaussian noise, contrast and brightness adjustments, rotations, and zooming, produced seven variants per image. Three machine learning models鈥擟NN, AdaBoost, and ResNet鈥攚ere evaluated, with the 850 nm light source yielding the best results: 87.5% of predictions fell within Zone A of the Clarke Error Grid. Findings support the potential of this approach for non-invasive glucose monitoring.
  Department: Computer Science
  Supervisor: Dr. Maria Valero
   |  | 
  
  

• GPR-132 Hyperparameter Optimization in Neural Network Using Binary Search Algorithm (PhD Research) by , , 
  Abstract: Hyperparameter searching is a crucial process for every neural network training. However, this process is notably time-consuming due to the vast number of possible combinations and the influence these hyperparameters have on each other. The common approach is using grid search to exhaust all the options, which is computationally very expensive. In this research, we propose a new algorithm for this problem that is inspired by binary search and returns a significant improvement in time efficiency.
  Department: Computer Science
  Supervisor: Dr Kazi Aminul Islam
   |  | 
  
  

• GPR-142 Optimization of Fixed Time in Round Robin Scheduling using Clustering Algorithms (PhD Research) by 
  Abstract: This project introduces a method to optimize the fixed time in Round Robin scheduling using unsupervised clustering, specifically DBSCAN. Traditionally, fixed time is chosen arbitrarily, often leading to inefficiencies like increased waiting times and frequent context switches. Our approach leverages DBSCAN to identify clusters of processes based on arrival and burst times, as well as to detect outliers that may need unique fixed times. This adaptive, data-driven adjustment has demonstrated improved performance over traditional methods, reducing waiting time, minimizing context switches, and enhancing system throughput. Simulations confirmed the effectiveness of this approach, especially in datasets with outlier processes, where DBSCAN performed exceptionally well.
  Department: Computer Science
  Supervisor: Dr. Dan Lo
   |  | 
  
  

• GPR-148 A study of different real-time robotic applications (PhD Research) by 
  Abstract: Real-time operating systems (RTOS) are widely used in various robotic applications such as path planning and obstacle avoidance, which require real-time communication and interaction with the environment, posing significant challenges for RTOS design.In this paper, we will first explore different robotic control and decision-making applications based on RTOS. Then, we will study the implementations of several widely employed RTOS frameworks. Finally, we will analyze how different RTOS implementations impact overall system performance and discuss the advantages and limitations of these RTOS frameworks based on previous research.
  Department: Computer Science
  Supervisor: Dr. Dan Lo
   | 
  
  

• GPR-151 Compassionate Digital Assistant: Anchor (PhD Research) by , 
  Abstract: Mental health support is crucial, but access to professional care can be limited by cost, availability, and social stigma. Digital solutions, particularly chatbots, offer an accessible and scalable approach to providing mental health support. However, current chatbot solutions may not always reflect the diversity of users' emotional experiences, and they may lack the specialized domain knowledge and adaptability required for effective mental health counseling. This research project aims to address these challenges by developing a compassionate AI digital assistant that can use specialized natural language processing (NLP) models to provide empathetic and targeted responses based on the nature of the user's query. Unlike traditional chatbots that rely on fine-tuning large language models, our approach is to build conversation graphs for specific mental health scenarios, which can maintain historical context and adapt in real-time to the user's emotional state. The motivation for this project stems from the growing mental health crisis and the projected shortage of mental health professionals in the coming years. As advancements in digital technology and the industrial economy have contributed to increased mental health challenges, the demand for mental health care is expected to grow significantly, outpacing the available workforce. By creating a digital assistant capable of providing emotional support and evidence-based guidance, this research aims to help fill the anticipated gap in mental health services and improve access to mental health resources for those in need.
  Department: Computer Science
  Supervisor: Dr. Md Abdullah Al Hafiz Khan
   |  | 
  
  

• GPR-155 Integration of Quantum Natural Language Processing (QNLP) with Neo4j LLM Knowledge Graphs for Enhanced NLP Tasks (PhD Research) by 
  Abstract: This study investigates the integration of Quantum Natural Language Processing (QNLP) with Neo4j LLM Knowledge Graphs (KGs) to enhance natural language understanding tasks. By leveraging quantum circuit simulations, we aim to improve the probabilistic interpretation of relationships between entities. Our preliminary findings suggest that QNLP offers deeper insights compared to traditional NLP methods, particularly in modeling complex entity relationships. This approach also addresses significant limitations in Neo4j-based Large Language Model (LLM) Graph Databases, such as handling high dimensional relationships and capturing semantic nuances. The integration of QNLP into Neo4j refines relationship modeling and enhances performance in tasks like entity extraction and knowledge inference, paving the way for more advanced and context-aware NLP applications.
  Department: Computer Science
  Supervisor: Dr. Dan Lo
   | 
  
  

• GPR-161 Meta-Reinforcement Learning with Discrete World Models for Adaptive Load Balancing (PhD Research) by 
  Abstract: We present a novel integration of the RL^2 meta-reinforcement learning algorithm with discrete world models, employing the DreamerV3 architecture, to enhance load balancing in operating systems. This integration allows for rapid adaptation to dynamic workload distributions with minimal retraining. In experiments using the Park load balancing environment, our approach outperformed the traditional AC3 algorithm in both standard and adaptive trials. Additionally, it exhibited strong resilience to catastrophic forgetting, maintaining high performance despite continuous variations in workload distribution and size. These results demonstrate the effectiveness of combining recurrent policy networks with discrete world models, offering a significant advancement in meta-learning capabilities for dynamic operating system environments. This work has important implications for improving resource management and performance in modern operating systems, addressing the challenges posed by increasingly dynamic and heterogeneous workloads.
  Department: Computer Science
  Supervisor: Dr. Dan Lo
   |  | 
  
  

• GPR-185 A Multimodal Approach to Quiz Generation: Leveraging RAG Models for Educational Assessments (PhD Research) by 
  Abstract: Crafting quiz questions that effectively assess students鈥� understanding of lectures and course materials, such as textbooks, poses significant challenges. Recent AI-based quiz generation efforts have predominantly concentrated on static resources, like textbooks and slides, often overlooking the dynamic and interactive elements of live lectures鈥攃ontextual cues, discussions, and interactions鈥攖hat contribute to the learning experience. In this work, we propose a Retrieval-Augmented Generation (RAG) model that processes multimodal inputs by combining text, audio, and video to produce quizzes that capture a fuller context. Our method incorporates Whisper for audio transcription and utilizes a Large Vision-Language Model (LVLM) to extract essential visual data from lecture videos. By integrating both spoken and visual elements, our model generates quizzes that more closely represent the lecture environment. We evaluate the model鈥檚 impact on quiz relevance, diversity, and engagement, showing that this multimodal approach fosters a more dynamic and immersive learning experience. Performance metrics, including hit rate and mean reciprocal rank (MRR), are used to assess question relevance and accuracy. A high hit rate indicates the model鈥檚 reliability in producing pertinent questions, while MRR highlights ranking quality, demonstrating the prompt appearance of relevant questions. Strong results in these metrics confirm our model鈥檚 effectiveness, though current limitations include challenges in handling abstract concepts absent in the lecture material鈥攁 gap we aim to bridge in future developments by integrating external knowledge sources.
  Department: Computer Science
  Supervisor: Dr. Nasrin Dehbozorgi
   | 
  
  

• GPR-187 Deep Learning Models for Protein-Protein Binding Affinity Prediction (PhD Research) by Chen, Lingtao
  Abstract: Binding affinity (BA) prediction is important for drug discovery and protein engineering. It seeks to understand the interaction strength between proteins and their ligands (or proteins). This information assists in the design of proteins with enhanced or novel functions, as well as understanding the molecular mechanisms of drug action. This paper presents the development and comparative analysis of two deep learning models, a convolutional neural network (CNN) and a transformer model. Many variants of models in this research were developed using TensorFlow. One model that utilizes ProteinBERT was developed using PyTorch. The CNN model captures local sequence features effectively, while the Transformer model leverages self-attention mechanisms to learn long-range dependencies within the sequences. Protein sequences are the inputs for the models. The sequences are processed using various encoders, like One-hot encoding, Sequence-Statistics-Content, and Position Specific Scoring Matrix. The predicted outputs are Gibbs free energy changes, a key indicator of binding affinity. From this study, both the CNN and transformer models can achieve the same level of accuracy under different conditions. For the CNN model, it can handle full data without sacrificing performance, but it takes much more time to preprocess the features from the protein sequences. The transformer model can achieve the same level of accuracy as the CNN model with no big predictive errors for each protein, but it requires the model to run on less data, which removes some rarely long protein sequences. This study emphasizes the potential of advanced deep learning architectures to enhance the predictive strengths of binding affinity models.
  Department: Computer Science
  Supervisor: Dr. Yixin Xie
   |  | 
  
  

• GPR-2212 Explainable Multi-Label Classification Framework for Behavioral Health Based on Domain Concepts (PhD Research) by Nweke, Francis E, Azmee, Abm Adnan
  Abstract: Behavioral health, which covers mental health, lifestyle choices, addictions, and crises, poses serious issues in the community. Thus, appropriately analyzing and classifying behavioral health data is crucial for making informed healthcare decisions. Traditional deep learning and natural language processing approaches struggle to effectively identify behavioral health issues because the data is unstructured, complex, and lacks sufficient context. Furthermore, subject matter experts must be consulted to ensure effective identification. In this work, we proposed a deep learning-based framework consisting of several modules: A) domain concept encoder converts the keywords and their evidence types to vectors, which were predefined by a subject matter expert; B) the semantic representation encoder (SRE) is trained on the vectors to learn the relationship between them; C) transformed-based feature learner is an advanced learner that extracts feature embeddings from documents and generates attention weights since it has more context given the incorporated relationship weights; D) The behavioral health multilabel classifier utilizes feature embeddings to classify a document into one or more behavioral health classes; and E) The LLM-enabled explainer provides explanations based on attention weights and classifications. Our proposed framework outperformed state-of-the-art models in multilabel behavioral health case classification while also providing explanations for each classification. Which is crucial in behavioral health analysis.
  Department: Computer Science
  Supervisor: Dr. Md Abdullah Al Hafiz Khan (PhD Advisor), Collaborators: Dr. Yong Pei, Dr. Dominic Thomas, Dr. Monica Nandan
   | 
  
  

• GPR-2238 Tech Guru: A Domain Specific LLM for Tech. Industry (PhD Research) by Nweke, Francis E, Vu, Long, Jha, Nitin
  Abstract: This project focuses on developing a domain-specific chatbot tailored for the tech industry. The chatbot utilizes articles sourced from blogs written by developers and engineers at leading companies such as Google and NVIDIA. Titles and content from these articles are extracted to form a question-answer dataset, with the titles acting as questions and the article content serving as answers. To refine the questions, we implemented a custom method to format the dataset to follow Alpaca format. The resulting question-answer pairs are then used to fine-tune a language model, adapting it to the specialized domain of the tech industry. Following this, the model undergoes rigorous evaluation to ensure its accuracy and relevance, and iterative improvements are made based on performance metrics. The final product is deployed as a chatbot capable of handling complex queries in the tech space, offering valuable support to developers and engineers.
  Department: Computer Science
  Supervisor: Dr. Ramazan Aygun
   | 
  
  

• GPR-233 Human-Assisted AI for Detecting Mental Health Indicators in Social Media (PhD Research) by , , 
  Abstract: Mental health is essential to overall well-being, and mental illness includes conditions that affect a person鈥檚 psychological health, causing significant distress and limiting daily functioning. With advancements in technology, social media has become a platform where individuals openly share their emotions and thoughts, offering a unique window into their psychological states. However, traditional machine learning models struggle to interpret social media data's wide range of linguistic nuances. To analyze this data effectively, collaboration with human experts is crucial. This study proposes an innovative human-AI teaming framework that integrates human expertise with artificial intelligence (AI) to address these challenges. Our framework leverages multi-dimensional data along with expert feedback to identify factors contributing to mental illness. Through extensive testing on Reddit data, our model demonstrates a 9% improvement in performance over the state-of-the-art model, underscoring its efficacy and impact.
  Department: Computer Science
  Supervisor: Dr. Md Abdullah Al Hafiz Khan (PhD Advisor)
   | 

• GPR-6126 Utilizing ML techniques for a Quantum Augmented HTTP Protocol (PhD Research) by 
  Abstract: Over the past decade, several small-scale quantum key distribution (QKD) networks have been implemented worldwide. However, achieving scalable, large-scale quantum networks relies on advancements in quantum repeaters, channels, memories, and network protocols. To enhance the security of current networks while utilizing available quantum technologies, integrating classical networks with quantum elements appears to be the next logical step. In this study, we propose modifications to the HTTP protocol's data packet structure, adjustments to end-to-end encryption methods, and optimized bandwidth distribution between quantum and classical channels for high-traffic network routes.
  Department: Computer Science
  Supervisor: Dr. Abhishek Parakh (KSU), Dr. Mahadevan Subramaniam (University of Nebraska Omaha)
   | 
  
  

• GPR-6128 Joint Encryption and Error Correction for Quantum Communication (PhD Research) by 
  Abstract: Secure quantum networks are foundational for developing a quantum internet in the future. However, current age quantum channels are prone to noise, which can introduce errors in transmitted data. Traditionally, error correction is applied to the message separately, after encryption, resulting in additional overhead that should be minimized. In response, we propose a unified approach that combines encryption and error correction into a single process. This work represents an initial effort to integrate these functions for secure quantum communication by integrating the Calderbank-Shor-Steane (CSS) Quantum Error Correction (QECC) code with the three-stage secure quantum communication protocol. Additionally, the protocol supports the transmission of arbitrary qubits from sender to receiver, making it adaptable for general use.
  Department: Computer Science
  Supervisor: Dr. Abhishek Parakh, Dr. Mahadevan Subramaniam
   |  |