Search:
Match:
1 results

Analysis

This paper addresses a significant gap in current world models by incorporating emotional understanding. It argues that emotion is crucial for accurate reasoning and decision-making, and demonstrates this through experiments. The proposed Large Emotional World Model (LEWM) and the Emotion-Why-How (EWH) dataset are key contributions, enabling the model to predict both future states and emotional transitions. This work has implications for more human-like AI and improved performance in social interaction tasks.
Reference

LEWM more accurately predicts emotion-driven social behaviors while maintaining comparable performance to general world models on basic tasks.