BlockPuzzle - A Challenge in Physical Reasoning and Generalization for Robot Learning
This work addresses the problem of sparse reward settings in reinforcement learning for robots, but it is incremental as it only represents a first step in solving harder tasks and transferring knowledge.
The authors tackled the challenge of physical reasoning puzzles in robot learning by proposing a novel task framework with simple rules, and they successfully solved several environments using curricula and imitation learning methods.
In this work we propose a novel task framework under which a variety of physical reasoning puzzles can be constructed using very simple rules. Under sparse reward settings, most of these tasks can be very challenging for a reinforcement learning agent to learn. We build several simple environments with this task framework in Mujoco and OpenAI gym and attempt to solve them. We are able to solve the environments by designing curricula to guide the agent in learning and using imitation learning methods to transfer knowledge from a simpler environment. This is only a first step for the task framework, and further research on how to solve the harder tasks and transfer knowledge between tasks is needed.