Extracting Location Descriptions from Disaster-Related Messages using Geo-Knowledge-Guided GPT Models
March 11, 2025 12:00 pm (Central Time)
Abstract
During natural disasters, people share urgent information on social media and chat groups, often including critical location descriptions. Accurately extracting these location descriptions can help responders reach victims faster. Traditional named entity recognition (NER) struggles with complex location descriptions, but recent geo-knowledge-guided GPT models have shown promise. In this presentation I will describe my Community Champion project and its contributions to the Spatial AI Challenge with its refined dataset of 1,000 Hurricane Harvey tweets with labeled location descriptions. A Jupyter Notebook with example code for processing the dataset and integrating geo-knowledge with OpenAI's GPT API will be presented. Researchers can use this resource to refine prompts, improve location extraction methods, or experiment with open-source LLMs.
Speakers
Yingjie Hu
University at Buffalo
Yingjie Hu is an Associate Professor in the Department of Geography at the University at Buffalo (UB). He is also an Adjunct Professor in the Department of Computer Science and Engineering, a Faculty Member of the UB Center for Geological and Climate Hazards, and an Affiliate Faculty of the UB AI and Data Science Institute. His major research area is geographic information science (GIScience), and more specifically geospatial artificial intelligence (GeoAI) and spatial data science. He leads the GeoAI Lab@UB, a research group focusing on integrating geospatial data, spatial analysis, and AI methods to understand human-environment interactions and to address some of the challenges confronting our society, such as those related to natural disasters, public health, and ecosystem conservation.