UDALM: Unsupervised Domain Adaptation through Language Modeling
This addresses domain adaptation for language models, providing a sample-efficient method for downstream tasks, though it appears incremental.
The paper tackles unsupervised domain adaptation for pretrained language models by introducing UDALM, a fine-tuning procedure using mixed classification and Masked Language Model loss, which achieves 91.74% accuracy on Amazon Reviews Sentiment dataset, a 1.11% absolute improvement over state-of-the-art.
In this work we explore Unsupervised Domain Adaptation (UDA) of pretrained language models for downstream tasks. We introduce UDALM, a fine-tuning procedure, using a mixed classification and Masked Language Model loss, that can adapt to the target domain distribution in a robust and sample efficient manner. Our experiments show that performance of models trained with the mixed loss scales with the amount of available target data and the mixed loss can be effectively used as a stopping criterion during UDA training. Furthermore, we discuss the relationship between A-distance and the target error and explore some limitations of the Domain Adversarial Training approach. Our method is evaluated on twelve domain pairs of the Amazon Reviews Sentiment dataset, yielding $91.74\%$ accuracy, which is an $1.11\%$ absolute improvement over the state-of-the-art.