Collecting data for training dialog systems can be extremely expensive due to the involvement of human participants and need for extensive annotation. Especially in document-grounded dialog systems, human experts need to carefully read the unstructured documents to answer the users' questions. As a result, existing document-grounded dialog datasets are relatively small-scale and obstruct the effective training of dialogue systems. In this paper, we propose an automatic data augmentation technique grounded on documents through a generative dialogue model. The dialogue model consists of a user bot and agent bot that can synthesize diverse dialogues given an input document, which are then used to train a downstream model. When supplementing the original dataset, our method achieves significant improvement over traditional data augmentation methods. We also achieve great performance in the low-resource setting.
@article{arxiv.2112.08342,
title = {DG2: Data Augmentation Through Document Grounded Dialogue Generation},
author = {Qingyang Wu and Song Feng and Derek Chen and Sachindra Joshi and Luis A. Lastras and Zhou Yu},
journal= {arXiv preprint arXiv:2112.08342},
year = {2021}
}