Learning from omission

McDowell, B; Goodman, ND

McDowell, B (reprint author), Duolingo, Pittsburgh, PA 15206 USA.

57TH ANNUAL MEETING OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS (ACL 2019), 2019; (): 619

Abstract

Pragmatic reasoning allows humans to go beyond the literal meaning when interpreting language in context. Previous work has shown that such reasoning can improve the performance of already-trained language understanding systems. Here, we explore whether pragmatic reasoning during training can improve the quality of learned meanings. Our experiments on reference game data show that end-to-end pragmatic training produces more accurate utterance interpretation models, especially when data is sparse and language is complex.

Download PDF


Full Text Link