Breaking NLI Systems with Sentences that Require Simple Lexical Inferences

Max Glockner, Vered Shwartz, Yoav Goldberg


Abstract
We create a new NLI test set that shows the deficiency of state-of-the-art models in inferences that require lexical and world knowledge. The new examples are simpler than the SNLI test set, containing sentences that differ by at most one word from sentences in the training set. Yet, the performance on the new test set is substantially worse across systems trained on SNLI, demonstrating that these systems are limited in their generalization ability, failing to capture many simple inferences.
Anthology ID:
P18-2103
Volume:
Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers)
Month:
July
Year:
2018
Address:
Melbourne, Australia
Editors:
Iryna Gurevych, Yusuke Miyao
Venue:
ACL
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
650–655
Language:
URL:
https://aclanthology.org/P18-2103
DOI:
10.18653/v1/P18-2103
Bibkey:
Cite (ACL):
Max Glockner, Vered Shwartz, and Yoav Goldberg. 2018. Breaking NLI Systems with Sentences that Require Simple Lexical Inferences. In Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), pages 650–655, Melbourne, Australia. Association for Computational Linguistics.
Cite (Informal):
Breaking NLI Systems with Sentences that Require Simple Lexical Inferences (Glockner et al., ACL 2018)
Copy Citation:
PDF:
https://aclanthology.org/P18-2103.pdf
Presentation:
 P18-2103.Presentation.pdf
Video:
 https://aclanthology.org/P18-2103.mp4
Code
 BIU-NLP/Breaking_NLI +  additional community code
Data
MultiNLISNLI