Termsslug: reward-shaping

Reward Shaping

奖励塑形

Introducing intermediate rewards or constraints to guide a model's learning trajectory so that stage-wise behavior aligns with t

Curated Intelligence

1 items · Page 1/1
新智元 (微信公众号)T2·Models
58 pts
CUHK and Microsoft Propose EchoChat to Curb Error Propagation in Empathetic Reasoning

To address error propagation across multi-stage reasoning in empathetic spoken dialogue, researchers from CUHK and Microsoft introduce EchoChat, a unified framework that organizes perception, theory-of-mind inference, strategy planning and speech response generation, constrains inter-stage dependencies with cascaded gated rewards, and applies step-level credit assignment for targeted updates. Accepted at NeurIPS 2026, the work reframes empathy as a decomposable cognitive process rather than a direct input-to-response mapping.