Use of Large Language Models to Improve Transportation Services
This project aims to leverage Large Language Models (LLMs) to enhance the analysis of public complaints and suggestions related to transportation systems. By processing feedback from multiple agencies, this study seeks to cluster and analyze common concerns, aiding agencies in aligning their services with public demands and safety needs. The project addresses equity, safety, and sustainability by identifying complaint patterns, especially in underserved areas, to inform responsive infrastructure policies. An open-source LLM model will be developed to safeguard privacy while enabling data-driven improvements in transportation services.
Language
- English
Project
- Status: Active
-
Contract Numbers:
69A3552348303
-
Sponsor Organizations:
Sustainable Mobility and Accessibility Regional Transportation Equity Research Center
Morgan State University
Baltimore, MD United StatesOffice of the Assistant Secretary for Research and Technology
University Transportation Centers Program
Department of Transportation
Washington, DC United States 20590 -
Managing Organizations:
Sustainable Mobility and Accessibility Regional Transportation Equity Research Center
Morgan State University
Baltimore, MD United States -
Performing Organizations:
University of Pittsburgh
Benedum Engineering Hall
Pittsburgh, PA United States 15261 -
Principal Investigators:
Khazanovich, Lev
Stevanovic, Aleksandar
- Start Date: 20240901
- Expected Completion Date: 20250901
- Actual Completion Date: 0
- USDOT Program: University Transportation Centers
Subject/Index Terms
- TRT Terms: Artificial intelligence; Cluster analysis; Customer satisfaction; Information processing; Language
- Subject Areas: Data and Information Technology; Planning and Forecasting; Transportation (General);
Filing Info
- Accession Number: 01935540
- Record Type: Research project
- Source Agency: Sustainable Mobility and Accessibility Regional Transportation Equity Research Center
- Contract Numbers: 69A3552348303
- Files: UTC, RIP
- Created Date: Oct 30 2024 2:52PM