AtmosERC Models Dialogue-Level Affective Atmosphere for Emotion Recognition
July 30, 2026
AtmosERC utilizes a relation-aware graph extractor to model conversational graphs over utterances and speakers. This approach estimates dialogue-level affective priors to guide sequential emotion prediction or provide verbalized prompt-level cues for LLM-based emotion recognition.
HOW THIS AFFECTS YOU
●
researcherYou can use these affective priors to improve LLM-based emotion recognition through prompt engineering.