GALAXY: A Generative Pre-trained Model for Task-Oriented Dialog with Semi-Supervised Learning and Explicit Policy Injection
This addresses the challenge of improving dialog policy in task-oriented systems for applications like virtual assistants, though it is incremental as it builds on existing pre-training methods.
The paper tackles the problem of enhancing task-oriented dialog systems by proposing GALAXY, a pre-trained model that explicitly learns dialog policy through semi-supervised learning, achieving state-of-the-art results with improvements of 2.5, 5.3, and 5.5 points on benchmark datasets.
Pre-trained models have proved to be powerful in enhancing task-oriented dialog systems. However, current pre-training methods mainly focus on enhancing dialog understanding and generation tasks while neglecting the exploitation of dialog policy. In this paper, we propose GALAXY, a novel pre-trained dialog model that explicitly learns dialog policy from limited labeled dialogs and large-scale unlabeled dialog corpora via semi-supervised learning. Specifically, we introduce a dialog act prediction task for policy optimization during pre-training and employ a consistency regularization term to refine the learned representation with the help of unlabeled dialogs. We also implement a gating mechanism to weigh suitable unlabeled dialog samples. Empirical results show that GALAXY substantially improves the performance of task-oriented dialog systems, and achieves new state-of-the-art results on benchmark datasets: In-Car, MultiWOZ2.0 and MultiWOZ2.1, improving their end-to-end combined scores by 2.5, 5.3 and 5.5 points, respectively. We also show that GALAXY has a stronger few-shot ability than existing models under various low-resource settings.