TY  - RPRT
TI  - ReF-HIL: Shaping the Critic around Human Action Neighborhoods for Efficient Human-in-the-Loop Reinforcement Learning
AU  - Shaoyin Luo
AU  - Song Wang
AU  - Shibo Xia
AU  - Tianle Zhang
AU  - Zhaowei Liang
AU  - Guanghui Shen
AU  - Bin Wang
AU  - Dan Wu
PY  - 2026
UR  - https://arxiv.org/abs/2609.37131
ID  - 2609.37131
ER  -