Latent Intention Dialogue Models

TL;DR

This paper introduces a Latent Intention Dialogue Model (LIDM) that uses discrete latent variables to capture dialogue intentions, improving goal-oriented dialogue generation through neural variational inference and reinforcement learning.

Contribution

The paper presents a novel LIDM framework that models dialogue intentions with discrete latent variables, enhancing naturalness and effectiveness in goal-oriented conversations.

Findings

LIDM outperforms existing benchmarks in corpus-based evaluations.

LIDM achieves higher scores in human evaluations.

Discrete latent variables effectively capture dialogue intentions.

Abstract

Developing a dialogue agent that is capable of making autonomous decisions and communicating by natural language is one of the long-term goals of machine learning research. Traditional approaches either rely on hand-crafting a small state-action set for applying reinforcement learning that is not scalable or constructing deterministic models for learning dialogue sentences that fail to capture natural conversational variability. In this paper, we propose a Latent Intention Dialogue Model (LIDM) that employs a discrete latent variable to learn underlying dialogue intentions in the framework of neural variational inference. In a goal-oriented dialogue scenario, these latent intentions can be interpreted as actions guiding the generation of machine responses, which can be further refined autonomously by reinforcement learning. The experimental evaluation of LIDM shows that the model out-performs published benchmarks for both corpus-based and human evaluation, demonstrating the effectiveness of discrete latent variable models for learning goal-oriented dialogues.