Evaluation Examples are not Equally Informative: How should that change NLP Leaderboards?

Pedro Rodriguez, Joe Barrow, Alexander Miserlis Hoyle, John P. Lalor, Robin Jia, Jordan Boyd-Graber


Abstract
Leaderboards are widely used in NLP and push the field forward. While leaderboards are a straightforward ranking of NLP models, this simplicity can mask nuances in evaluation items (examples) and subjects (NLP models). Rather than replace leaderboards, we advocate a re-imagining so that they better highlight if and where progress is made. Building on educational testing, we create a Bayesian leaderboard model where latent subject skill and latent item difficulty predict correct responses. Using this model, we analyze the ranking reliability of leaderboards. Afterwards, we show the model can guide what to annotate, identify annotation errors, detect overfitting, and identify informative examples. We conclude with recommendations for future benchmark tasks.
Anthology ID:
2021.acl-long.346
Volume:
Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers)
Month:
August
Year:
2021
Address:
Online
Venues:
ACL | IJCNLP
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
4486–4503
Language:
URL:
https://aclanthology.org/2021.acl-long.346
DOI:
10.18653/v1/2021.acl-long.346
Bibkey:
Copy Citation:
PDF:
https://aclanthology.org/2021.acl-long.346.pdf
Optional supplementary material:
 2021.acl-long.346.OptionalSupplementaryMaterial.zip