TY  - RPRT
TI  - Automating the Refinement of Reinforcement Learning Specifications
AU  - Tanmay Ambadkar
AU  - Đorđe Žikelić
AU  - Abhinav Verma
PY  - 2026
UR  - https://arxiv.org/abs/2512.01047
ID  - 2512.01047
ER  -