Emotion Transition

Emotion transition research focuses on understanding and modeling how emotions change over time, particularly within conversational or expressive contexts. Current work explores this through various approaches, including reinforcement learning for optimizing emotional support conversations, weakly supervised learning for generating realistic 3D gestures reflecting emotional shifts, and the use of large language models and neural networks to generate natural language descriptions of emotional changes in speech or music. This research is significant for advancing human-computer interaction, improving mental health support systems, and creating more emotionally intelligent and engaging virtual experiences.