Bayesian robot control handles sudden environment shifts safely

Original: BAPR: Bayesian amnesic piecewise-robust reinforcement learning for non-stationary continuous control

Source: arxiv.org ↗