TOD-BERT: Pre-trained Natural Language Understanding For Task-oriented Dialogue

Wu Chien-sheng, Hoi Steven, Socher Richard, Xiong Caiming. Arxiv 2020

[Paper]    
Applications BERT Few Shot Language Modeling Masked Language Model Model Architecture Pretraining Methods Reinforcement Learning Training Techniques

The underlying difference of linguistic patterns between general text and task-oriented dialogue makes existing pre-trained language models less useful in practice. In this work, we unify nine human-human and multi-turn task-oriented dialogue datasets for language modeling. To better model dialogue behavior during pre-training, we incorporate user and system tokens into the masked language modeling. We propose a contrastive objective function to simulate the response selection task. Our pre-trained task-oriented dialogue BERT (TOD-BERT) outperforms strong baselines like BERT on four downstream task-oriented dialogue applications, including intention recognition, dialogue state tracking, dialogue act prediction, and response selection. We also show that TOD-BERT has a stronger few-shot ability that can mitigate the data scarcity problem for task-oriented dialogue.

Similar Work