ILDAE: Instance-Level Difficulty Analysis of Evaluation Data

Neeraj Varshney, Swaroop Mishra, Chitta Baral


Abstract
Knowledge of difficulty level of questions helps a teacher in several ways, such as estimating students’ potential quickly by asking carefully selected questions and improving quality of examination by modifying trivial and hard questions. Can we extract such benefits of instance difficulty in Natural Language Processing? To this end, we conduct Instance-Level Difficulty Analysis of Evaluation data (ILDAE) in a large-scale setup of 23 datasets and demonstrate its five novel applications: 1) conducting efficient-yet-accurate evaluations with fewer instances saving computational cost and time, 2) improving quality of existing evaluation datasets by repairing erroneous and trivial instances, 3) selecting the best model based on application requirements, 4) analyzing dataset characteristics for guiding future data creation, 5) estimating Out-of-Domain performance reliably. Comprehensive experiments for these applications lead to several interesting results, such as evaluation using just 5% instances (selected via ILDAE) achieves as high as 0.93 Kendall correlation with evaluation using complete dataset and computing weighted accuracy using difficulty scores leads to 5.2% higher correlation with Out-of-Domain performance. We release the difficulty scores and hope our work will encourage research in this important yet understudied field of leveraging instance difficulty in evaluations.
Anthology ID:
2022.acl-long.240
Volume:
Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long 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:
3412–3425
Language:
URL:
https://aclanthology.org/2022.acl-long.240
DOI:
10.18653/v1/2022.acl-long.240
Bibkey:
Cite (ACL):
Neeraj Varshney, Swaroop Mishra, and Chitta Baral. 2022. ILDAE: Instance-Level Difficulty Analysis of Evaluation Data. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 3412–3425, Dublin, Ireland. Association for Computational Linguistics.
Cite (Informal):
ILDAE: Instance-Level Difficulty Analysis of Evaluation Data (Varshney et al., ACL 2022)
Copy Citation:
PDF:
https://aclanthology.org/2022.acl-long.240.pdf
Software:
 2022.acl-long.240.software.zip
Video:
 https://aclanthology.org/2022.acl-long.240.mp4
Code
 nrjvarshney/ildae
Data
GLUEMultiNLIPAWSQuaRTzQuaRelSNLISWAGWinoGrande