Data Augmentation for Conversational AI

arXiv:2309.04739v2 Announce Type: replace
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 …

abstract arxiv augmentation challenge collection context conversational conversational ai cs.cl cs.ir data data collection dialogue domains information labor languages limitations low making queries systems them training training data type