Simplified Factual Statement Extractor

This page provides source code for a system for extracting simplified factual statements from complex sentences. The system utilizes a set of Tregex tree searching rules to identify, in phrase structure trees for input sentences, instances of various constructions (e.g., relative clauses, participial phrases). It then extracts simplified declarative sentences from these constructions.

It is described further in the following paper: Michael Heilman and Noah A. Smith. 2010. Extracting Simplified Statements for Factual Question Generation. Proc. of the 3rd Workshop on Question Generation. [PDF]

Source Code

The following compressed archive includes the Java source code, a compiled JAR file, various dependencies such as the Stanford Parser, and some documentation: FactualStatementExtractor-06-15-2010.tar.gz.