Paper: Hypothesis and Evidence Extraction from Full-Text Scientific Journal Articles

ACL ID W11-0217
Title Hypothesis and Evidence Extraction from Full-Text Scientific Journal Articles
Venue Workshop on Biomedical Natural Language Processing
Session
Year 2011
Authors

Increasingly, as full-text scientific papers are becoming available, scientific queries have shifted from looking for facts to looking for arguments. Researchers want to know when their colleagues are proposing theories, out- lining evidentiary relations, or explaining dis- crepancies. We show here that sentence-level annotation with the CISP schema adapts well to a corpus of biomedical articles, and we present preliminary results arguing that the CISP schema is uniquely suited to recovering common types of scientific arguments about hypotheses, explanations, and evidence.