Invariant Language Modeling

Maxime Peyrard, Sarvjeet Ghotra, Martin Josifoski, Vidhan Agarwal, Barun Patra, Dean Carignan, Emre Kiciman, Saurabh Tiwary, Robert West


Abstract
Modern pretrained language models are critical components of NLP pipelines. Yet, they suffer from spurious correlations, poor out-of-domain generalization, and biases.Inspired by recent progress in causal machine learning, in particular the invariant risk minimization (IRM) paradigm, we propose invariant language modeling, a framework for learning invariant representations that generalize better across multiple environments. In particular, we adapt a game-theoretic implementation of IRM (IRM-games) to language models, where the invariance emerges from a specific training schedule in which all the environments compete to optimize their own environment-specific loss by updating subsets of the model in a round-robin fashion.We focused on controlled experiments to precisely demonstrate the ability of our method to (i) remove structured noise, (ii) ignore specific spurious correlations without affecting global performance, and (iii) achieve better out-of-domain generalization.These benefits come with a negligible computational overhead compared to standard training, do not require changing the local loss, and can be applied to any language model. We believe this framework is promising to help mitigate spurious correlations and biases in language models.
Anthology ID:
2022.emnlp-main.387
Volume:
Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing
Month:
December
Year:
2022
Address:
Abu Dhabi, United Arab Emirates
Editors:
Yoav Goldberg, Zornitsa Kozareva, Yue Zhang
Venue:
EMNLP
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
5728–5743
Language:
URL:
https://aclanthology.org/2022.emnlp-main.387
DOI:
10.18653/v1/2022.emnlp-main.387
Bibkey:
Cite (ACL):
Maxime Peyrard, Sarvjeet Ghotra, Martin Josifoski, Vidhan Agarwal, Barun Patra, Dean Carignan, Emre Kiciman, Saurabh Tiwary, and Robert West. 2022. Invariant Language Modeling. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 5728–5743, Abu Dhabi, United Arab Emirates. Association for Computational Linguistics.
Cite (Informal):
Invariant Language Modeling (Peyrard et al., EMNLP 2022)
Copy Citation:
PDF:
https://aclanthology.org/2022.emnlp-main.387.pdf