Analyzing Geographic Questions Using Embedding-based Topic Modeling

Recently, open-domain question-answering systems have achieved tremendous progress because of developments in large language models (LLMs), and successfully been applied to (QA) systems, or Chatbots. However, there has little question answering the geographic domain. Existing research domain relies heavily on rule-based semantic parsing approaches using few data. To develop intelligent GeoQA agents, it is crucial build QA upon datasets that reflect real users’ needs regarding studies analyzed questions corpora Microsoft MAchine Reading Comprehension (MS MARCO), comprising real-world user queries from Bing terms structural similarity, which does not discover interests. Therefore, we aimed analyze location-related MS MARCO based group similar into a cluster, utilize results interests Using sentence-embedding-based topic modeling approach cluster semantically questions, obtained could gather documents single cluster. Furthermore, discovered latent topics within collection guide practical relevant questions.