A Tri-Partite Neural Document Language Model for Semantic Information Retrieval

Abstract

Previous work in information retrieval have shown that using evidence, such as concepts and relations, from external knowledge sources could enhance the retrieval performance. Recently, deep neural approaches have emerged as state-of-the art models for capturing word semantics. This paper presents a new tri-partite neural document language framework that leverages explicit knowledge to jointly constrain word, concept, and document learning representations to tackle a number of issues including polysemy and granularity mismatch. We show the effectiveness of the framework in various IR tasks.