TY  - RPRT
TI  - Reward function compression facilitates goal-dependent reinforcement learning
AU  - Gaia Molinaro
AU  - Anne G. E. Collins
PY  - 2026
UR  - https://arxiv.org/abs/2509.06810
ID  - 2509.06810
ER  -