Simulated Annealing for Emotional Dialogue Systems

TL;DR

This paper introduces a search-based dialogue generation method using simulated annealing to explicitly incorporate specific emotions, improving emotional accuracy without sacrificing response quality.

Contribution

It proposes a novel search-based approach with simulated annealing for emotion-controlled dialogue generation, addressing limitations of previous input-based methods.

Findings

12% improvement in emotion accuracy over state-of-the-art

Maintains BLEU scores, indicating preserved generation quality

Effective in expressing specific emotions in dialogue responses

Abstract

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).