Annotation artifacts in natural language inference data

Suchin Gururangan, Swabha Swayamdipta, Omer Levy, Roy Schwartz, Samuel R. Bowman, Noah A. Smith

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Large-scale datasets for natural language inference are created by presenting crowd workers with a sentence (premise), and asking them to generate three new sentences (hypotheses) that it entails, contradicts, or is logically neutral with respect to. We show that, in a significant portion of such data, this protocol leaves clues that make it possible to identify the label by looking only at the hypothesis, without observing the premise. Specifically, we show that a simple text categorization model can correctly classify the hypothesis alone in about 67% of SNLI (Bowman et al., 2015) and 53% of MultiNLI (Williams et al., 2018). Our analysis reveals that specific linguistic phenomena such as negation and vagueness are highly correlated with certain inference classes. Our findings suggest that the success of natural language inference models to date has been overestimated, and that the task remains a hard open problem.

Original languageEnglish
Title of host publicationShort Papers
PublisherAssociation for Computational Linguistics (ACL)
Pages107-112
Number of pages6
ISBN (Electronic)9781948087292
StatePublished - 2018
Externally publishedYes
Event2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, NAACL HLT 2018 - New Orleans, United States
Duration: 1 Jun 20186 Jun 2018

Publication series

NameNAACL HLT 2018 - 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies - Proceedings of the Conference
Volume2

Conference

Conference2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, NAACL HLT 2018
Country/TerritoryUnited States
CityNew Orleans
Period1/06/186/06/18

Fingerprint

Dive into the research topics of 'Annotation artifacts in natural language inference data'. Together they form a unique fingerprint.

Cite this