Complex Question Answering Via Reasoning Across Multiple Text Passages

W. Lam

Automatic answering natural language queries or questions issued by users can facilitate the development of a wide range of intelligent applications. We intend to investigate a practical setting in which there exists a large collection of text documents. The aim is to infer and find the answer from the document collection given a question issued by the user. We investigate a new framework for answering complex natural language queries and questions capable of conducting reasoning and integrating evidence derived across multiple text snippets or passages from different documents.