EmPO: Emotion Grounding for Empathetic Response Generation through Preference Optimization
This work addresses the problem of enhancing emotional intelligence in AI chatbots for more engaging human-machine interactions, representing an incremental improvement by applying existing preference optimization methods to a specific domain.
The paper tackled the challenge of generating empathetic responses in conversational agents by aligning large language models with theory-driven preference datasets based on emotion grounding, using preference optimization to improve empathy metrics like diff-Epitome and BERTscore while retaining generalization performance on benchmarks like MMLU.
Empathetic response generation is a desirable aspect of conversational agents, crucial for facilitating engaging and emotionally intelligent multi-turn conversations between humans and machines. Leveraging large language models for this task has shown promising results, yet challenges persist in ensuring both the empathetic quality of the responses and retention of the generalization performance of the models. We propose a novel approach where we construct theory-driven preference datasets based on emotion grounding and use them to align LLMs with preference optimization algorithms to address these challenges. To evaluate empathetic response generation, we employ the EmpatheticDialogues dataset, assessing empathy with the diff-Epitome and BERTscore metrics and with multi-dimensional human evaluation. Additionally, we measure diversity and emotional valence using feature-based methods. We also evaluate the impact of training on the generalization performance using the MMLU benchmark and tasks from the Open LLM Leaderboard. The results show that LLMs can be aligned for empathetic response generation by preference optimization while retaining their general performance and that emotion grounding can guide preference dataset creation. We make all datasets, source code, and models publicly available. https://github.com/justtherightsize/empo