Leave no Trace: Learning to Reset for Safe and Autonomous Reinforcement Learning
This addresses the need for autonomous and safe reinforcement learning in robotics by reducing human intervention, though it is incremental as it builds on existing reset policy ideas.
The paper tackles the problem of requiring manual resets in reinforcement learning for non-reversible tasks by proposing a method that learns both forward and reset policies, resulting in reduced manual resets and unsafe actions, with experiments showing significant reductions in these metrics.
Deep reinforcement learning algorithms can learn complex behavioral skills, but real-world application of these methods requires a large amount of experience to be collected by the agent. In practical settings, such as robotics, this involves repeatedly attempting a task, resetting the environment between each attempt. However, not all tasks are easily or automatically reversible. In practice, this learning process requires extensive human intervention. In this work, we propose an autonomous method for safe and efficient reinforcement learning that simultaneously learns a forward and reset policy, with the reset policy resetting the environment for a subsequent attempt. By learning a value function for the reset policy, we can automatically determine when the forward policy is about to enter a non-reversible state, providing for uncertainty-aware safety aborts. Our experiments illustrate that proper use of the reset policy can greatly reduce the number of manual resets required to learn a task, can reduce the number of unsafe actions that lead to non-reversible states, and can automatically induce a curriculum.