Automatic Single-Document Key Fact Extraction from Newswire Articles

This paper addresses the problem of extracting the most important facts from a news article. Our approach uses syntactic, semantic, and general statistical features to identify the most important sentences in a document. The importance of the individual features is estimated using generalized iterative scaling methods trained on an annotated newswire corpus. The performance of our approach is evaluated against 300 unseen news articles and shows that use of these features results in statistically significant improvements over a provenly robust baseline, as measured using metrics such as precision, recall and ROUGE.