Emotion Entanglement and Bayesian Inference for Multi-Dimensional Emotion Understanding 事件

Emotion Entanglement and Bayesian Inference for Multi-Dimensional Emotion Understanding arXiv:2604.00819v2 Announce Type: replace Abstract: Understanding emotions in natural language is inherently a multi-dimensional reasoning problem, where multiple affective signals interact through context, interpersonal relations, and situational cues. However, most existing emotion understanding benchmarks rely on short texts and predefined emotion labels, reducing this process to independent label predict