AI

Overcoming the Lack of Parallel Data in Sentence Compression

Abstract

A subset of the described data (10,000 sentence & extracted headlines pairs, with source URL and annotations) is available for download.