Back-Training excels Self-Training at Unsupervised Domain Adaptation of Question Generation and Passage Retrieval

Abstract

In this work, we introduce back-training, an alternative to self-training for unsupervised domain adaptation (UDA) from source to target domain. While self-training generates synthetic training data where natural inputs are aligned with noisy outputs, back-training results in natural outputs aligned with noisy inputs. This significantly reduces the gap between the target domain and synthetic data distribution, and reduces model overfitting to the source domain. We run UDA experiments on question generation and passage retrieval from the \textit{Natural Questions} domain to machine learning and biomedical domains. We find that back-training vastly outperforms self-training by a mean improvement of 7.8 BLEU-4 points on generation, and 17.6\% top-20 retrieval accuracy across both domains. We further propose consistency filters to remove low-quality synthetic data before training. We also release a new domain-adaptation dataset- \textit{MLQuestions} containing 35K unaligned questions, 50K unaligned passages, and 3K aligned question-passage pairs.

arXiv commentEMNLP 2021

References in corpus (7)

• Unsupervised Statistical Machine Translation
• Question Answering and Question Generation as Dual Tasks
• Style Transfer as Unsupervised Machine Translation
• A Joint Model for Question Answering and Question Generation
• Learning to Answer by Learning to Ask: Getting the Best of GPT-2 and BERT Worlds
• End-to-End QA on COVID-19: Domain Adaptation with Synthetic Training
• Towards Automatic Generation of Questions from Long Answers