TY  - RPRT
TI  - A Contractive Feedback Semantics for Reinforcement Learning
AU  - Zuyuan Zhang
PY  - 2026
UR  - https://arxiv.org/abs/2605.24759
ID  - 2605.24759
ER  -