TY - RPRT TI - Model-Agnostic Solutions for Deep Reinforcement Learning in Non-Ergodic Contexts AU - Bert Verbruggen AU - Arne Vanhoyweghen AU - Vincent Ginis PY - 2026 UR - https://arxiv.org/abs/2601.08726 ID - 2601.08726 ER -