Undergraduate Research Fellowship
The SDS Undergraduate Research Fellowship offers a paid opportunity for students of all majors to engage in impactful research under the guidance of UTSA faculty. The application form includes short-answer questions and a faculty reference check.
Students should review the faculty research projects listed below and indicate their preferences. While we aim to match students with projects that align with their interests, placements depend on project needs and availability, and specific matches cannot be guaranteed.
Please Note: The faculty proposals listed here are subject to selection based on student fit and program requirements. Not all proposals may be matched or selected for this fellowship.
Projects
Preliminary Study of AI-Based Personalized Music Intervention for Alleviating Anxiety and Depression
This project explores the feasibility of using artificial intelligence to generate personalized music interventions aimed at reducing mental distress, such as anxiety and depression. The objective is to develop and evaluate an AI-driven music generation model tailored to the psychological needs of individual subjects. We will collaborate with the UTSA Psychology Department to recruit participants and assess the effectiveness of personalized AI-generated music in reducing mental health symptoms. The key areas of focus include AI model development, music personalization algorithms, and psychological impact assessment.
KubeLLM: Collaborative LLM Agents for Distributed Edge Computing Systems
In this project, we propose KubeLLM, a framework that integrates Large Language Models (LLMs) with Kubernetes-based orchestration to enhance the management of distributed edge computing systems. KubeLLM embeds lightweight LLM agents within edge nodes to enable automated monitoring, diagnosis, and resolution of configuration issues. This approach aims to proactively address silent failures, minimize unexpected disruptions, and mitigate security vulnerabilities, thereby ensuring robust and reliable edge computing environments
Semi-Supervised Learning Approaches for Defect Detection in Nuclear Materials Fabrication
This project from the UTSA Vision & Artificial Intelligence Lab in Computer Science is part of a larger collaboration with the UTSA Extreme Environment Materials Lab in Physics, focused on the application of novel AI methods to materials fabrication. The project will explore unsupervised and semi-supervised learning approaches to detecting defects in images of fabricated nuclear materials. Analysis of these materials and defects will help to inform the fabrication process for our collaborators.
EfficientAI: Compressive Sensing and AI Integration for Low-Power, Energy-Harvesting IoT Systems in Data Science Applications
This project focuses on implementing compressed sensing and AI algorithms on battery-less IoT device efficiently to reduce the need of sampling and storing large amount of signals for health monitoring applications while addressing the intrinsic weakness of battery-less IoT devices. One of the potential data applications is to design and implement an AI/ML algorithm that can automatically discriminate life-threatening ventricular arrhythmias (VAs) (i.e., binary classification: VAs or non-VAs) from sparse (labeled one-channel intracardiac electrograms recordings while being able to be deployed and run on the given microcontroller platform.
Advanced Data Science Techniques for Enhancing Biological Dataset Usability
Modern biomedical science has been revolutionized by the advancement of experimental techniques that can collect data at an unprecedented scale and resolution. As a result, biomedical research is increasingly relying on computational methods to extract and analyze these data to reveal the hidden meaning and help make critical healthcare decisions. My lab is dedicated to developing data-analytical methods and tools to make complex biological data more understandable and useful. In this SDS Undergraduate Research Fellowship Program, you will have the opportunity to learn some advanced data science techniques developed in the lab, make improvements, and apply them to some interesting biological applications provided by the instructor (such as cancer outcome prediction and drug resistance prediction), or applying to some applications of your interest.
Improve Mobility of Patients with Smart Wheelchairs
This project aims to develop a smart wheelchair to assist individuals with reduced mobility. The key area of research includes implement of camera to image acquiring, application of AI for object detection on Internet of Things, motion control and path planning, and status update along the path using RFID.
Intelligent Smart-Phone Based Assistive Driving Solution for Power Wheelchairs
Power wheelchairs (PWCs) have been widely utilized to improve the independence of people with disabilities. Although PWCs usually provide convenient joystick-type interfaces that can be easily operated with hands, for people with severe cognitive, motion, or sensory issues, it is still a difficult task to operate PWCs to their fullest extent for an extended period of time. To alleviate the operation difficulties and ease the burden and stress of disabled individuals, this project focuses on the design and development of a cost-effective, smartphone-based Intelligent Assistive Driving system for Power Wheelchair (PWC IA-Driving).
Optimizing Distributed Deep Neural Network (DNN) Inference for Low Latency and Energy Efficiency on Resource-Constrained Edge Devices
This research project aims to develop and evaluate novel methods for optimizing distributed deep neural network (DNN) inference to achieve low latency and energy efficiency on resource-constrained edge devices. Current approaches for distributed DNN inference typically involve splitting neural network layers or operations across multiple edge devices. However, these configurations are often pre-determined and fail to adapt to fluctuations in network latency and bandwidth, resulting in performance degradation and increased energy consumption.
Quantifying representation bias in the inputs and outputs of LLM on multimodal medical data (tabular and textual)
This project aims to study representation bias and healthcare disparity in medical data and the LLM (small scale) built over it. The main focus of this project will be hands-on data analysis, developing a small-scale LLM, and understanding its behaviors. Students will be given access to real datasets from the medical domain and a pre-trained small-scale LLM as a starting point. Three undergraduate students will be preferred and the project will also function with two students under a limited budget.
Quality Assurance of Radiation Treatments for Cancer Patients using Machine Learning
In collaboration with Mays Cancer Center (UT Health SA), I am working to develop a machine learning (ML)-based QA tool for radiotherapy. This tool aims to be more robust in handling data irregularities, while simultaneously reducing the clinical burden associated with radiotherapy evaluation processes. The main objective of this undergraduate research project will be to evaluate the pros and cons of using ML for radiotherapy QA. The key focus area will be data science/ML and medical physics (field of radiation therapy).
Machine Learning Application for Rupture Status Assessments of Intracranial Aneurysms
In this research, we aim to apply various machine learning (ML) algorithms to assess the rupture status of intracranial aneurysms (IA). Specifically, we will evaluate seven widely used ML algorithms: multivariate logistic regression (LR), support vector machine (SVM), random forest (RF), extreme gradient boosting (XGBoost), multi-layer perceptron neural network (MLPNN), Bayesian additive regression trees (BART), and deep learning techniques. We will assess the performance of all these methods in terms of prediction accuracy and the area under the receiver operating characteristic curve (AUC) on the testing samples. Next, we will compare the LIME and SHAP algorithms to analyze how each feature contributes to the final prediction for the best-performing model. Finally, we will provide recommendations to facilitate the clinical translation of these models.
Tracing Ecological Changes During the Early Cretaceous Using Thin Section Analysis
About 120 million years ago, marine ecosystems had to adapt to major environmental changes that resulted in oceanic anoxia. The objectives of this project are to (1) trace changes in the ecology of these ecosystems based on thin sections analysis, and (2) identify what environmental parameter forces the ecosystems to adapt based on the review of existing geochemical dataset. Students will have hands-on experience in the fields of microscopic analysis of rock samples, data analysis using the language R, and presentation of scientific results.
Training Large Language Models at the Speed of Light
As large language models (LLMs)—the technology behind chatbots, translation tools, and other AI applications—become increasingly complex, the amount of computation required to train them has skyrocketed. Traditional computer processors often struggle to handle this demand efficiently. This project aims to explore silicon photonics, a groundbreaking technology that uses light, not electricity, to process and transmit data. Light is much faster and more energy-efficient than electrical signals, making silicon photonics a potential game-changer for training large AI models.
