TY  - RPRT
TI  - Counteractive RL: Rethinking Core Principles for Efficient and Scalable Deep Reinforcement Learning
AU  - Ezgi Korkmaz
PY  - 2026
UR  - https://arxiv.org/abs/2603.15871
ID  - 2603.15871
ER  -