AuGPT: Auxiliary Tasks and Data Augmentation for End-To-End Dialogue with Pre-Trained Language Models

TL;DR

AuGPT enhances end-to-end dialogue models by introducing modified training objectives and extensive data augmentation, significantly improving performance on the MultiWOZ dataset through better knowledge grounding and diversity.

Contribution

The paper proposes new training objectives and employs large-scale data augmentation techniques to improve dialogue modeling with pre-trained language models.

Findings

Outperforms baseline on MultiWOZ dataset

Achieves competitive results with state-of-the-art models

Improves diversity and knowledge grounding in dialogue generation

Abstract

Attention-based pre-trained language models such as GPT-2 brought considerable progress to end-to-end dialogue modelling. However, they also present considerable risks for task-oriented dialogue, such as lack of knowledge grounding or diversity. To address these issues, we introduce modified training objectives for language model finetuning, and we employ massive data augmentation via back-translation to increase the diversity of the training data. We further examine the possibilities of combining data from multiples sources to improve performance on the target dataset. We carefully evaluate our contributions with both human and automatic methods. Our model substantially outperforms the baseline on the MultiWOZ data and shows competitive performance with state of the art in both automatic and human evaluation.