TY  - RPRT
TI  - Specifying Reward Functions for RL Without Environment Sampling
AU  - Stephane Hatgis-Kessell
AU  - W. Bradley Knox
AU  - Emma Brunskill
PY  - 2026
UR  - https://arxiv.org/abs/2609.15544
ID  - 2609.15544
ER  -