RE-LLM: Refining Empathetic Speech-LLM Responses by Integrating Emotion Nuance
This work addresses the need for deeper emotional engagement in human-AI interaction, offering incremental improvements in empathetic response generation for conversational AI systems.
The paper tackled the problem of limited emotion nuance in empathetic AI responses by proposing RE-LLM, a speech-LLM integrating dimensional emotion embeddings and auxiliary learning, resulting in statistically significant gains in empathy metrics, such as improving the Emotional Reaction score by up to 14.79% and Exploration scores by up to 139.28% across datasets.
With generative AI advancing, empathy in human-AI interaction is essential. While prior work focuses on emotional reflection, emotional exploration, key to deeper engagement, remains overlooked. Existing LLMs rely on text which captures limited emotion nuances. To address this, we propose RE-LLM, a speech-LLM integrating dimensional emotion embeddings and auxiliary learning. Experiments show statistically significant gains in empathy metrics across three datasets. RE-LLM relatively improves the Emotional Reaction score by 14.79% and 6.76% compared to text-only and speech-LLM baselines on ESD. Notably, it raises the Exploration score by 35.42% and 3.91% on IEMOCAP, 139.28% and 9.83% on ESD, and 60.95% and 22.64% on MSP-PODCAST. It also boosts unweighted accuracy by 5.4% on IEMOCAP, 2.3% on ESD, and 6.9% on MSP-PODCAST in speech emotion recognition. These results highlight the enriched emotional understanding and improved empathetic response generation of RE-LLM.