TY  - RPRT
TI  - ACCRUE: Continual Off-Policy Reinforcement Learning for Robot Foundation Models
AU  - Changyeon Kim
AU  - Kimin Lee
AU  - Yashraj Narang
AU  - Yijie Guo
PY  - 2026
UR  - https://arxiv.org/abs/2610.18596
ID  - 2610.18596
ER  -