Large Language Models for Spatiotemporal Event Recommendation

Keywords: GeoAI, LLM, recommendation, Large Language Model
Abstract Type: Paper Abstract

Abstract

Recommending similar events that come with geospatial and temporal information is an essential task in modern geographic information retrieval tools. This paper introduces a novel retrieval-reranking framework leveraging Large Language Models (LLMs) to enhance the spatiotemporal and semantic associated searching and recommendation. This framework uses advanced large models to address the limitations of traditional manual curation in terms of high labor cost and lack of scalability and bridges critical information across modalities. We apply the proposed framework to surface similar climate change events described in news articles and web posts. We hope that by automatically linking relevant events, this framework can better assist critical incident curators and the general public in gaining an enhanced understanding of climate change and its impact on different communities. The proposed recommendation framework can be applied to a wide range of similar search and recommendation tasks dealing with geospatial and temporal data, using the power of Generative AI in GeoAI.