A 4.5B-parameter multimodal model trained on a new 130K reasoning-trajectory dataset with a multi-task reinforcement learning recipe reports near-commercial performance across eight affective tasks.
Emotional artificial intelligence in education: A systematic review and meta- analysis,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.HC 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
OneEmo: A Unified Multimodal Reasoning Model for Emotion Perception, Understanding, and Interaction
A 4.5B-parameter multimodal model trained on a new 130K reasoning-trajectory dataset with a multi-task reinforcement learning recipe reports near-commercial performance across eight affective tasks.