Experience Summary
Student Projects
The UTSA School of Data Science Summer Immersion Experience offered students from Tec de Monterrey and Instituto Tecnológico Autónomo de México (ITAM) a chance to expand their professional networks. Students collaborated with peers and faculty on various projects, gaining practical experience and receiving career coaching and guidance.
| Name | Description | Poster |
|---|---|---|
| Smart City: Managing Electric Vehicle Demand in Smart Grids | The objective was to efficiently match EV customers to electric feeders in San Antonio, ensuring each feeder meets demand while minimizing the total distance between them. | Smart City Research Poster |
| Collaborative multi-agent language models for software development | Software development involves collaboration across various skills, including design, coding, and testing. This project introduces a novel approach for using LLMs in backend development through agent-based methods and advanced prompting while exploring cutting-edge evaluation techniques. | |
| Mitigating Dialect Biases in Privacy Policy Question-Answering Systems through Collaborative Agent-based Language Models |
This is a project to question the performance disparities in privacy policy question-answering systems across different English Dialects and to create a collaborative multiagent-based solution to mitigate these biases, promoting fairness in automated systems. | Mitigating Dialect Biases Research Poster |
| Optimizing Pedestrian Paths for Maximum Shadow Coverage: Balancing Efficiency and Distance |
Pedestrian navigation apps typically optimize for distance but often overlook user preferences, such as climate considerations. Walking through cities involves more than just finding the shortest path; pedestrians also seek to maximize their well-being by, for example, walking on shaded paths. | Optimizing Pedestrian Paths Research Poster |
| Unity-Based Real-Time 3D Human Pose Estimation and Shape Modeling |
This project integrates pose data from Google MediaPipe Pose Estimation with a SMPL humanoid model in Unity. The multi-threaded tracking system, built in Python, sends 3D pose results to Unity for modeling and visualization, with the model, webcam input, and game running on separate threads. |
|
| 3D Human Motion Capture by Combining Kinect and Vicon Data | The widespread use of VR in gaming, training, and simulation has led to advancements and safety risks like collisions. Current VR safety systems often rely on a single sensor type, limiting their effectiveness. This project enhances safety by integrating motion capture data from both Vicon cameras and Kinect sensors. | 3D Human Motion Capture Research Poster |
| Towards Equitable Machine Learning Driven Housing Price Prediction |
This project provides an accurate prediction of house pricing in San Antonio, and works to improve the regressor model fairness on ethnicity and race criterion. |
|
| Game-Theoretic Multi-Agent Large Language Models for Auction Simulation and Trading |
This project aims to use the simulation platform from the referenced paper to create a digital twin with a multi-agent model for radio spectrum auctions in telecommunications. The goal is to develop a regulatory sandbox to improve decision-making on rules and optimal auction design. |
Students' Cultural and Networking Experience
During the 4-week program, students from Tec de Monterrey and ITAM had opportunities to participate in various networking events and field trips to immerse themselves in UTSA's and San Antonio, Texas culture.
