Design Guidelines for Developing Systems for Dialogue System Competitions

Kazunori Komatani, Ryu Takeda, Keisuke Nakashima, Mikio Nakano

Abstract: Because dialogue system development involves a variety of factors and requires multifaceted consideration, design guidelines for such development would be helpful. Although a neural-based approach can be used, it requires a vast amount of dialogue data, which would take too much effort to collect in the case of a system for a specific, fixed-length dialogue. Furthermore, the system design should explicitly consider errors in automatic speech recognition and language understanding, because they degrade the user impression and are inevitable when the system talks with general users. Accordingly, we propose design guidelines for developing such dialogue systems. Systems developed with the aid of these guidelines took first place in two dialogue system competitions: the situation track of the second Dialogue System Live Competition and a pre-preliminary contest of the Dialogue Robot Competition. Our proposed design guidelines are to: (1) make the system take initiative, (2) prevent dialogue flows from relying too much on user utterances, and (3) include in utterances that the system understands what the user said. We describe details and examples for the systems designed for each of the two competitions.