Proceedings of the ACL 2003 workshop on Multilingual summarization and question answering -

Abstract

Automatic summarization and question answering aim at producing a concise, condensed representation of the key information content in an information source for a particular user and task. Interest in automatic summarization and question answering continues to grow, motivated by the explosion of on-line information sources and advances in natural language processing and information retrieval. In fact, various forms of automatic summarization and question answering will undoubtedly be indispensable given the massive information universes that lie ahead in the 21st century.Summarization and question answering involves the extraction or generation of text snippets to fulfill some user needs. Rule-based or statistical-based summarization and QA systems have shown promising results in the TREC QA-tracks, NTCIR QAC, and NIST DUC; it is, however, very difficult to find good evaluation functions or rules that work well across domains or in all questions because there are many system parameters that must be carefully tuned in order to achieve good system performance. In consequence, various machine learning (ML) techniques have recently been applied to summarization and QA systems.