Multi-armed bandits for resource efficient, online optimization of language model pre-training: the use case of dynamic masking

Inigo Urteaga, Moulay Zaidane Draidia, Tomer Lancewicki, Shahram Khadivi


Abstract
We design and evaluate a Bayesian optimization framework for resource efficient pre-training of Transformer-based language models (TLMs). TLM pre-training requires high computational resources and introduces many unresolved design choices, such as selecting its pre-training hyperparameters.We propose a multi-armed bandit framework for the sequential selection of pre-training hyperparameters, aimed at optimizing language model performance, in a resource efficient manner. We design a Thompson sampling algorithm, with a surrogate Gaussian process reward model of the Masked Language Model (MLM) pre-training objective, for its sequential minimization. Instead of MLM pre-training with fixed masking probabilities, the proposed Gaussian process-based Thompson sampling (GP-TS) accelerates pre-training by sequentially selecting masking hyperparameters that improve performance. We empirically demonstrate how GP-TS pre-trains language models efficiently, i.e., it achieves lower MLM loss in fewer epochs, across a variety of settings. In addition, GP-TS pre-trained TLMs attain competitive downstream performance, while avoiding expensive hyperparameter grid search. GP-TS provides an interactive framework for efficient and optimized TLM pre-training that, by circumventing costly hyperparameter selection, enables substantial computational savings.
Anthology ID:
2023.findings-acl.675
Volume:
Findings of the Association for Computational Linguistics: ACL 2023
Month:
July
Year:
2023
Address:
Toronto, Canada
Editors:
Anna Rogers, Jordan Boyd-Graber, Naoaki Okazaki
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
10609–10627
Language:
URL:
https://aclanthology.org/2023.findings-acl.675
DOI:
10.18653/v1/2023.findings-acl.675
Bibkey:
Cite (ACL):
Inigo Urteaga, Moulay Zaidane Draidia, Tomer Lancewicki, and Shahram Khadivi. 2023. Multi-armed bandits for resource efficient, online optimization of language model pre-training: the use case of dynamic masking. In Findings of the Association for Computational Linguistics: ACL 2023, pages 10609–10627, Toronto, Canada. Association for Computational Linguistics.
Cite (Informal):
Multi-armed bandits for resource efficient, online optimization of language model pre-training: the use case of dynamic masking (Urteaga et al., Findings 2023)
Copy Citation:
PDF:
https://preview.aclanthology.org/improve-issue-templates/2023.findings-acl.675.pdf