Topic-switch adapted Japanese Dialogue System based on PLATO-2

TL;DR

This paper develops a large-scale Japanese dialogue system based on PLATO-2 by creating a new dataset, Dialogue-Graph, and introduces a topic-switch algorithm to enhance user experience, achieving high evaluation scores.

Contribution

It introduces Dialogue-Graph, a large Japanese dialogue dataset, and a topic-switch algorithm, adapting PLATO-2 for Japanese and improving topic transition handling.

Findings

PLATO-JDS achieves high human evaluation scores close to maximum.

The topic-switch algorithm significantly improves user experience.

The system demonstrates effective dialogue generation in Japanese.

Abstract

Large-scale open-domain dialogue systems such as PLATO-2 have achieved state-of-the-art scores in both English and Chinese. However, little work explores whether such dialogue systems also work well in the Japanese language. In this work, we create a large-scale Japanese dialogue dataset, Dialogue-Graph, which contains 1.656 million dialogue data in a tree structure from News, TV subtitles, and Wikipedia corpus. Then, we train PLATO-2 using Dialogue-Graph to build a large-scale Japanese dialogue system, PLATO-JDS. In addition, to improve the PLATO-JDS in the topic switch issue, we introduce a topic-switch algorithm composed of a topic discriminator to switch to a new topic when user input differs from the previous topic. We evaluate the user experience by using our model with respect to four metrics, namely, coherence, informativeness, engagingness, and humanness. As a result, our proposed PLATO-JDS achieves an average score of 1.500 for the human evaluation with human-bot chat strategy, which is close to the maximum score of 2.000 and suggests the high-quality dialogue generation capability of PLATO-2 in Japanese. Furthermore, our proposed topic-switch algorithm achieves an average score of 1.767 and outperforms PLATO-JDS by 0.267, indicating its effectiveness in improving the user experience of our system.