Look Before You Leap: Safe Model-Based Reinforcement Learning with Human Intervention
This addresses safety challenges for real-world RL applications, offering an incremental improvement by integrating human oversight and model-based methods.
The paper tackled the problem of ensuring safety in deep reinforcement learning by proposing MBHI, a model-based framework that uses human-like blocking decisions and imagined trajectories to prevent catastrophic states, achieving better sample efficiency and fewer catastrophes than baselines.
Safety has become one of the main challenges of applying deep reinforcement learning to real world systems. Currently, the incorporation of external knowledge such as human oversight is the only means to prevent the agent from visiting the catastrophic state. In this paper, we propose MBHI, a novel framework for safe model-based reinforcement learning, which ensures safety in the state-level and can effectively avoid both "local" and "non-local" catastrophes. An ensemble of supervised learners are trained in MBHI to imitate human blocking decisions. Similar to human decision-making process, MBHI will roll out an imagined trajectory in the dynamics model before executing actions to the environment, and estimate its safety. When the imagination encounters a catastrophe, MBHI will block the current action and use an efficient MPC method to output a safety policy. We evaluate our method on several safety tasks, and the results show that MBHI achieved better performance in terms of sample efficiency and number of catastrophes compared to the baselines.