A Network-based End-to-End Trainable Task-oriented Dialogue System

TL;DR

This paper presents an end-to-end neural network-based dialogue system that simplifies development and data collection for task-oriented conversations, demonstrated effectively in a restaurant search domain.

Contribution

Introduces a neural network-based end-to-end trainable dialogue system and a novel dialogue data collection method using a pipe-lined Wizard-of-Oz framework.

Findings

System can converse naturally with humans.

Effective in helping users accomplish restaurant search tasks.

Simplifies development process without extensive handcrafted components.

Abstract

Teaching machines to accomplish tasks by conversing naturally with humans is challenging. Currently, developing task-oriented dialogue systems requires creating multiple components and typically this involves either a large amount of handcrafting, or acquiring costly labelled datasets to solve a statistical learning problem for each component. In this work we introduce a neural network-based text-in, text-out end-to-end trainable goal-oriented dialogue system along with a new way of collecting dialogue data based on a novel pipe-lined Wizard-of-Oz framework. This approach allows us to develop dialogue systems easily and without making too many assumptions about the task at hand. The results show that the model can converse with human subjects naturally whilst helping them to accomplish tasks in a restaurant search domain.