CoDA21: Evaluating Language Understanding Capabilities of NLP Models With Context-Definition Alignment

Lütfi Kerem Senel, Timo Schick, Hinrich Schuetze


Abstract
Pretrained language models (PLMs) have achieved superhuman performance on many benchmarks, creating a need for harder tasks. We introduce CoDA21 (Context Definition Alignment), a challenging benchmark that measures natural language understanding (NLU) capabilities of PLMs: Given a definition and a context each for k words, but not the words themselves, the task is to align the k definitions with the k contexts. CoDA21 requires a deep understanding of contexts and definitions, including complex inference and world knowledge. We find that there is a large gap between human and PLM performance, suggesting that CoDA21 measures an aspect of NLU that is not sufficiently covered in existing benchmarks.
Anthology ID:
2022.acl-short.92
Volume:
Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers)
Month:
May
Year:
2022
Address:
Dublin, Ireland
Editors:
Smaranda Muresan, Preslav Nakov, Aline Villavicencio
Venue:
ACL
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
815–824
Language:
URL:
https://aclanthology.org/2022.acl-short.92
DOI:
10.18653/v1/2022.acl-short.92
Bibkey:
Cite (ACL):
Lütfi Kerem Senel, Timo Schick, and Hinrich Schuetze. 2022. CoDA21: Evaluating Language Understanding Capabilities of NLP Models With Context-Definition Alignment. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), pages 815–824, Dublin, Ireland. Association for Computational Linguistics.
Cite (Informal):
CoDA21: Evaluating Language Understanding Capabilities of NLP Models With Context-Definition Alignment (Senel et al., ACL 2022)
Copy Citation:
PDF:
https://aclanthology.org/2022.acl-short.92.pdf
Video:
 https://aclanthology.org/2022.acl-short.92.mp4
Code
 lksenel/coda21