DFM: Dialogue Foundation Model for Universal Large-Scale Dialogue-Oriented Task Learning

TL;DR

This paper introduces DFM, a large-scale unified dialogue foundation model trained on a diverse dataset, achieving state-of-the-art performance across multiple dialogue tasks.

Contribution

The paper presents a novel unified framework and auxiliary self-supervised tasks for training a universal dialogue model on a large, diverse dataset.

Findings

DFM outperforms models of similar size on various dialogue tasks.

The use of auxiliary self-supervised tasks stabilizes training on diverse data.

DFM significantly extends the capabilities of unified dialogue models.

Abstract

Building a universal conversational agent has been a long-standing goal of the dialogue research community. Most previous works only focus on a small set of dialogue tasks. In this work, we aim to build a unified dialogue foundation model (DFM) which can be used to solve massive diverse dialogue tasks. To achieve this goal, a large-scale well-annotated dialogue dataset with rich task diversity (DialogZoo) is collected. We introduce a framework to unify all dialogue tasks and propose novel auxiliary self-supervised tasks to achieve stable training of DFM on the highly diverse large scale DialogZoo corpus. Experiments show that, compared with models of the same size, DFM can achieve state-of-the-art or competitive performance on very rich cross-domain downstream dialogue tasks. This demonstrates that DFM largely extends the ability of unified dialogue pre-trained model.