LILLIE: Information extraction and database integration using linguistics and learning-based algorithms

Ellery Smith, Dimitris Papadopoulos, Martin Braschler, Kurt Stockinger

Abstract: Highlights•A novel, generic method to extract open information triples from unstructured text.•Substantially outperforms state-of-the-art systems on CaRB and Re-OIE16 benchmarks.•Combines linguistics and learning-based methods to balance both precision and recall.•Refines triples with dependency tree rules from a high-recall learning-based engine.•Includes several augmentations to modify the generality and granularity of triples.