Paper: A New Entity Salience Task with Millions of Training Examples

ACL ID E14-4040
Title A New Entity Salience Task with Millions of Training Examples
Venue Annual Meeting of The European Chapter of The Association of Computational Linguistics
Session Main Conference
Year 2014
Authors

Although many NLP systems are moving toward entity-based processing, most still identify important phrases using classi- cal keyword-based approaches. To bridge this gap, we introduce the task of entity salience: assigning a relevance score to each entity in a document. We demon- strate how a labeled corpus for the task can be automatically generated from a cor- pus of documents and accompanying ab- stracts. We then show how a classifier with features derived from a standard NLP pipeline outperforms a strong baseline by 34%. Finally, we outline initial experi- ments on further improving accuracy by