Self-Supervised Dialogue Learning

TL;DR

This paper introduces a self-supervised learning task for dialogue systems that leverages utterance order to improve coherence, using a novel network and adversarial training to enhance dialogue quality.

Contribution

It proposes a new self-supervised task and network for capturing dialogue flow, improving coherence in both open-domain and task-oriented systems.

Findings

Achieves state-of-the-art results on OpenSubtitles dataset.

Improves dialogue coherence and relevance.

Applicable to both open-domain and task-oriented dialogues.

Abstract

The sequential order of utterances is often meaningful in coherent dialogues, and the order changes of utterances could lead to low-quality and incoherent conversations. We consider the order information as a crucial supervised signal for dialogue learning, which, however, has been neglected by many previous dialogue systems. Therefore, in this paper, we introduce a self-supervised learning task, inconsistent order detection, to explicitly capture the flow of conversation in dialogues. Given a sampled utterance pair triple, the task is to predict whether it is ordered or misordered. Then we propose a sampling-based self-supervised network SSN to perform the prediction with sampled triple references from previous dialogue history. Furthermore, we design a joint learning framework where SSN can guide the dialogue systems towards more coherent and relevant dialogue learning through adversarial training. We demonstrate that the proposed methods can be applied to both open-domain and task-oriented dialogue scenarios, and achieve the new state-of-the-art performance on the OpenSubtitiles and Movie-Ticket Booking datasets.