Task 1: Emotion Intelligence

The Emotion Intelligence Track aims to evaluate the emotional competence of spoken dialogue systems across five critical dimensions. These dimensions capture how well a system can perceive, interpret, express, and respond to human emotions in interactive scenarios


  • Emotion Recognition: Identifying emotions such as happy, sad, anger, fear, neutral, surprise, and disgust from speech serves as the foundation of emotional intelligence in spoken interactions.
  • Emotion Expression: Generating speech that conveys specific emotions, testing the model’s control over tone, pitch, and rhythm.
  • Empathy Response: Responding to the user’s emotional state with appropriate emotional tone and empathy, ensuring the response reflects understanding and support for the user’s feelings.
  • Emotion Causal Reasoning: Inferring the underlying causes of emotions in conversation, helping the model understand the user’s emotional context.
  • Emotion Dynamics Analysis: Detecting dynamic emotional shifts in the user’s speech, iden- tifying significant changes in emotional state throughout the conversation.