The project’s goal is to develop a software, which uses Large Language Models (LLMs) to predict changes in urban environments over time and space by analyzing multiple types of data.
It aims to integrate and analyze diverse spatiotemporal data sources, including population flows, traffic patterns, and building usage, to provide a comprehensive understanding and prediction of urban dynamics.
The software development involves learning from multimodal spatiotemporal data unique to urban settings, optimizing the model to infer long-term (diachronic) urban changes, and enhancing the model’s ability to provide explainable predictions while minimizing inaccuracies (hallucinations) through methods like Retrieval-Augmented Generation (RAG) integrated with GIS software.
Urban Large Language Models (LLM)
Development of Large Language Model-based Multimodal Urban Spatiotemporal Changes Prediction
Project Details
Subject
LLM-based Multimodal Spatiotemporal Urban Change Prediction
Funded by
Ministry of Education of the Republic of Korea
National Research Foundation of Korea
Period: