Simulated Chats for Building Dialog Systems: Learning to Generate Conversations from Instructions

TL;DR

This paper introduces a method using GPT-2 to simulate conversations for training dialog systems, reducing reliance on large annotated datasets and improving performance in low-resource scenarios.

Contribution

The authors propose a novel data creation strategy that employs GPT-2 to generate simulated dialogues from instructions, enhancing dialog system training with less annotated data.

Findings

Simulated data improves dialog system performance in low-resource settings.

Method reduces need for extensive crowd-annotated datasets.

Significant improvements observed on MultiWOZ and PersonaChat datasets.

Abstract

Popular dialog datasets such as MultiWOZ are created by providing crowd workers an instruction, expressed in natural language, that describes the task to be accomplished. Crowd workers play the role of a user and an agent to generate dialogs to accomplish tasks involving booking restaurant tables, calling a taxi etc. In this paper, we present a data creation strategy that uses the pre-trained language model, GPT2, to simulate the interaction between crowd workers by creating a user bot and an agent bot. We train the simulators using a smaller percentage of actual crowd-generated conversations and their corresponding instructions. We demonstrate that by using the simulated data, we achieve significant improvements in low-resource settings on two publicly available datasets - the MultiWOZ dataset and the Persona chat dataset.