Incorporation of corpus-specific semantic information into question answering context

ABSTRACT

In today's environment of information overload, Question Answering (QA) is a critically important research area for the Semantic Web. In order for humans to make effective use of the expansive information sources available to us, we require automated tools to help us make sense of large amounts of data. Within this framework, Question Context plays an important role. We define Question Context to be an semantic structure that can be used to enrich queries so that the user's information need is better represented. This paper describes the theoretical foundations of a novel approach that uses statistical language modeling techniques to create Question Context and to then integrate it into the Information Retrieval stage of QA. We base our approach on two established language modeling methods - the Aspect Model, which is the basis of Probabilistic Latent Semantic Analysis (PLSA) and Relevance-Based Language Models. Our approach proposes an Aspect-Based Relevance Language Model as the Question Context Model, and our methodology incorporates corpus-specific semantic concepts into the QA process. Words from the most heavily relevant aspects are then incorporated into the query. We present some interesting preliminary qualitative results that show the potential usefulness of the Question Context Model to both the first (IR) and second (Intelligent Information Processing) stages of QA.

REFERENCES

• 1: Dempster, A. P., Laird, N. M., and Rubin, D. B., "Maximum Likelihood from Incomplete Data via the EM Algorithm," Journal of the Royal Statistical Society, vol. 39, pp. 1--38, 1977.
• 2: L. Hirschman , R. Gaizauskas, Natural language question answering: the view from here, Natural Language Engineering, v.7 n.4, p.275-300, December 2001 [doi>10.1017/S1351324901002807]
• 3: Thomas Hofmann, Probabilistic latent semantic indexing, Proceedings of the 22nd annual international ACM SIGIR conference on Research and development in information retrieval, p.50-57, August 15-19, 1999, Berkeley, California, United States [doi>10.1145/312624.312649]
• 4: Victor Lavrenko , W. Bruce Croft, Relevance based language models, Proceedings of the 24th annual international ACM SIGIR conference on Research and development in information retrieval, p.120-127, September 2001, New Orleans, Louisiana, United States [doi>10.1145/383952.383972]
• 5: Christopher D. Manning , Prabhakar Raghavan , Hinrich Schtze, Introduction to Information Retrieval, Cambridge University Press, New York, NY, 2008
• 6: Strohman, T., Metzler, D., Turtle, H., and Croft, W. B., "Indri: A language model-based search engine for complex queries," in presented as a poster at the International Conference on Intelligence Analysis McLean, VA, 2005.
• 7: Ellen M. Voorhees , Donna K. Harman, TREC: Experiment and Evaluation in Information Retrieval (Digital Libraries and Electronic Publishing), The MIT Press, 2005
• 8: Voorhees, E. M., "Overview of the TREC 2006 Question Answering Track," in Online proceedings of 2006 Text Retrieval Conference, 2006.

INDEX TERMS

Primary Classification:
  H. Information Systems
  H.3 INFORMATION STORAGE AND RETRIEVAL
      H.3.1 Content Analysis and Indexing
          Subjects: Linguistic processing

Additional Classification:
  H. Information Systems
  H.3 INFORMATION STORAGE AND RETRIEVAL
      H.3.1 Content Analysis and Indexing
          Subjects: Indexing methods

General Terms:
Algorithms, Design, Measurement, Performance

Keywords:
information retrieval, language modeling, question answering, statistical language modeling