Data Augmentation for Conversational AI

TL;DR

This paper reviews data augmentation techniques for conversational AI, emphasizing their importance in low-resource settings, recent advances, evaluation methods, challenges, and future research directions.

Contribution

It provides a comprehensive overview of recent data augmentation approaches in conversational systems, highlighting new methods and evaluation paradigms.

Findings

Recent advances in conversation augmentation techniques

Open domain and task-oriented conversation generation methods

Discussion of challenges and future research directions

Abstract

Advancements in conversational systems have revolutionized information access, surpassing the limitations of single queries. However, developing dialogue systems requires a large amount of training data, which is a challenge in low-resource domains and languages. Traditional data collection methods like crowd-sourcing are labor-intensive and time-consuming, making them ineffective in this context. Data augmentation (DA) is an affective approach to alleviate the data scarcity problem in conversational systems. This tutorial provides a comprehensive and up-to-date overview of DA approaches in the context of conversational systems. It highlights recent advances in conversation augmentation, open domain and task-oriented conversation generation, and different paradigms of evaluating these models. We also discuss current challenges and future directions in order to help researchers and practitioners to further advance the field in this area.