Efficient Domain Adaptation of Language Models via Adaptive Tokenization

TL;DR

This paper introduces an adaptive tokenization method for domain adaptation of language models, achieving near domain-specific pretraining performance with less training time and smaller models.

Contribution

It proposes a novel tokenizer adaptation approach based on divergence in token distributions, reducing the need for extensive domain-specific pretraining.

Findings

Achieves >97% of domain-specific pretraining performance

Produces smaller models with less training and inference time

Faster adaptation using tokenizer adjustment compared to full pretraining

Abstract

Contextual embedding-based language models trained on large data sets, such as BERT and RoBERTa, provide strong performance across a wide range of tasks and are ubiquitous in modern NLP. It has been observed that fine-tuning these models on tasks involving data from domains different from that on which they were pretrained can lead to suboptimal performance. Recent work has explored approaches to adapt pretrained language models to new domains by incorporating additional pretraining using domain-specific corpora and task data. We propose an alternative approach for transferring pretrained language models to new domains by adapting their tokenizers. We show that domain-specific subword sequences can be efficiently determined directly from divergences in the conditional token distributions of the base and domain-specific corpora. In datasets from four disparate domains, we find adaptive tokenization on a pretrained RoBERTa model provides >97% of the performance benefits of domain specific pretraining. Our approach produces smaller models and less training and inference time than other approaches using tokenizer augmentation. While adaptive tokenization incurs a 6% increase in model parameters in our experimentation, due to the introduction of 10k new domain-specific tokens, our approach, using 64 vCPUs, is 72x faster than further pretraining the language model on domain-specific corpora on 8 TPUs.