Developing Offline Strategies for Answering Medical Questions

We describe ongoing developments on two offline strategies for automatically answering questions in the medical domain: one based on an analysis of the document structure, the other based on dependency parsing. We highlight differences with open domain question answering, and provide a preliminary evaluation of the current state of our strategies.

[1]  Eduard H. Hovy,et al.  Learning surface text patterns for a Question Answering System , 2002, ACL.

[2]  Ellen M. Voorhees,et al.  Overview of the TREC 2002 Question Answering Track , 2003, TREC.

[3]  Valentin Jijkoun,et al.  Answer Selection in a Multi-stream Open Domain Question Answering System , 2004, ECIR.

[4]  Marti A. Hearst.  Text Tiling: Segmenting Text into Multi-paragraph Subtopic Passages , 1997, CL.

[5]  Martin M. Soubbotin,et al.  Use of Patterns for Detection of Likely Answer Strings: A Systematic Approach , 2002, TREC.

[6]  Gilad Mishne,et al.  Making Stone Soup: Evaluating a Recall-Oriented Multi-stream Question Answering System for Dutch , 2004, CLEF.

[7]  Stephen Soderland,et al.  Learning Information Extraction Rules for Semi-Structured and Free Text , 1999, Machine Learning.

[8]  Gertjan van Noord,et al.  Alpino: Wide-coverage Computational Analysis of Dutch , 2000, CLIN.

[9]  Pierre Zweigenbaum.  Question answering in biomedicine , 2003 .

[10]  Salim Roukos,et al.  Automatic Derivation of Surface Text Patterns for a Maximum Entropy Based Question Answering System , 2003, NAACL.

[11]  Gilad Mishne,et al.  How frogs built the Berlin Wall , 2004 .

[12]  Eduard H. Hovy,et al.  Offline Strategies for Online Question Answering: Answering Questions Before They Are Asked , 2003, ACL.

[13]  Robert Malouf,et al.  Wide Coverage Parsing with Stochastic Attribute Value Grammars , 2004 .

[14]  Valentin Jijkoun,et al.  Information Extraction for Question Answering: Improving Recall Through Syntactic Patterns , 2004, COLING.