Every time I fire a conversational designer, the performance of the dialog system goes down

TL;DR

This paper demonstrates that incorporating explicit domain knowledge from conversational designers into neural dialogue systems significantly improves performance, especially with limited annotated data, by using a semi-logical rule-based system called CLINN.

Contribution

The paper introduces CLINN, a novel system that encodes explicit knowledge as semi-logical rules, and shows its effectiveness in enhancing neural dialogue systems with less annotated data.

Findings

Explicit domain knowledge reduces data requirements.

CLINN outperforms state-of-the-art neural systems on MultiWOZ.

Rules from designers improve dialogue system performance.

Abstract

Incorporating explicit domain knowledge into neural-based task-oriented dialogue systems is an effective way to reduce the need of large sets of annotated dialogues. In this paper, we investigate how the use of explicit domain knowledge of conversational designers affects the performance of neural-based dialogue systems. To support this investigation, we propose the Conversational-Logic-Injection-in-Neural-Network system (CLINN) where explicit knowledge is coded in semi-logical rules. By using CLINN, we evaluated semi-logical rules produced by a team of differently skilled conversational designers. We experimented with the Restaurant topic of the MultiWOZ dataset. Results show that external knowledge is extremely important for reducing the need of annotated examples for conversational systems. In fact, rules from conversational designers used in CLINN significantly outperform a state-of-the-art neural-based dialogue system.