Dialogue State Induction Using Neural Latent Variable Models

TL;DR

This paper introduces neural latent variable models for automatically inducing dialogue states from unlabeled customer service dialogues, improving dialogue system performance without manual labeling.

Contribution

The paper proposes two neural latent variable models for unsupervised dialogue state induction, reducing reliance on costly manual annotations in dialogue systems.

Findings

Models effectively identify meaningful dialogue slots.

Induced dialogue states improve dialogue system performance.

Unsupervised approach covers diverse domains.

Abstract

Dialogue state modules are a useful component in a task-oriented dialogue system. Traditional methods find dialogue states by manually labeling training corpora, upon which neural models are trained. However, the labeling process can be costly, slow, error-prone, and more importantly, cannot cover the vast range of domains in real-world dialogues for customer service. We propose the task of dialogue state induction, building two neural latent variable models that mine dialogue states automatically from unlabeled customer service dialogue records. Results show that the models can effectively find meaningful slots. In addition, equipped with induced dialogue states, a state-of-the-art dialogue system gives better performance compared with not using a dialogue state module.