Answer Credibility: A Language Modeling Approach to Answer Validation

Document Type


Publication Date

Summer 5-2009


Answer Validation is a topic of significant interest within the Question Answering community. In this paper, we propose the use of language modeling methodologies for Answer Validation, using corpus-based methods that do not require the use of external sources. Specifically, we propose a model for Answer Credibility which quantifies the reliability of a source document that contains a candidate answer and the Question’s Context Model.


North American Association for Computational Linguistics: Human Language Technologies conference proceedings are freely available from the ACL at http://www.aclweb.org/anthology/N/N09/N09-2040.pdf. Copyright © 2009 Association for Computational Linguistics.