Explicitly modeling emotions in dialogue generation has important applications, such as building empathetic personal companions. In this study, we consider the task of expressing a specific emotion for dialogue generation. Previous approaches take the emotion as an input signal, which may be ignored during inference. We instead propose a search-based emotional dialogue system by simulated annealing (SA). Specifically, we first define a scoring function that combines contextual coherence and emotional correctness. Then, SA iteratively edits a general response and searches for a sentence with a higher score, enforcing the presence of the desired emotion. We evaluate our system on the NLPCC2017 dataset. Our proposed method shows 12% improvements in emotion accuracy compared with the previous state-of-the-art method, without hurting the generation quality (measured by BLEU).
@article{arxiv.2109.10715,
title = {Simulated Annealing for Emotional Dialogue Systems},
author = {Chengzhang Dong and Chenyang Huang and Osmar Zaïane and Lili Mou},
journal= {arXiv preprint arXiv:2109.10715},
year = {2021}
}