Knowledge-grounded dialogue generation aims to mitigate the issue of text degeneration by incorporating external knowledge to supplement the context. However, the model often fails to internalize this information into responses in a human-like manner. Instead, it simply inserts segments of the provided knowledge into generic responses. As a result, the generated responses tend to be tedious, incoherent, and in lack of interactivity which means the degeneration problem is still unsolved. In this work, we first find that such copying-style degeneration is primarily due to the weak likelihood objective, which allows the model to "cheat" the objective by merely duplicating knowledge segments in a superficial pattern matching based on overlap. To overcome this challenge, we then propose a Multi-level Adaptive Contrastive Learning (MACL) framework that dynamically samples negative examples and subsequently penalizes degeneration behaviors at both the token-level and sequence-level. Extensive experiments on the WoW dataset demonstrate the effectiveness of our approach across various pre-trained models.
@article{arxiv.2310.08943,
title = {Multi-level Adaptive Contrastive Learning for Knowledge Internalization in Dialogue Generation},
author = {Chenxu Yang and Zheng Lin and Lanrui Wang and Chong Tian and Liang Pang and Jiangnan Li and Qirong Ho and Yanan Cao and Weiping Wang},
journal= {arXiv preprint arXiv:2310.08943},
year = {2023}
}