Enterprise Q&A System using Large Language Models (LLMs)

Transportation applications involve a wide array of data modalities, including maps, camera imagery, sensor outputs, tabular statistics, system logs, GPS traces, and spatial databases. Effectively reasoning over and integrating such heterogeneous data types necessitates the use of specialized agents capable of understanding and querying each modality. This project aims to develop multi-agent Large Language Model (LLM) platform to enhance reasoning capabilities within transportation domains. The focus areas include agent-based map interpretation and manipulation, spatially-aware query processing, natural language summarization and audio narration of traffic or spatial data, AI-driven assistance for accident reporting and transportation-related queries, and chatbot functionalities for internal enterprise datasets. Our target is to integrate the developed innovations and techniques into the Fanar GenAI platform.

Hamad Bin Khalifa University (HBKU)