Learning Real-World Robot Policies by Dreaming
This work addresses the problem of sample inefficiency in robotics for researchers and practitioners, offering a novel approach but with incremental improvements over existing world model methods.
The paper tackles the challenge of learning robot control policies from images by proposing a dreaming model that emulates realistic environment dynamics, enabling policy learning without extensive real-world interaction. The method successfully transfers learned policies to real-world robots, though no specific performance numbers are provided.
Learning to control robots directly based on images is a primary challenge in robotics. However, many existing reinforcement learning approaches require iteratively obtaining millions of robot samples to learn a policy, which can take significant time. In this paper, we focus on learning a realistic world model capturing the dynamics of scene changes conditioned on robot actions. Our dreaming model can emulate samples equivalent to a sequence of images from the actual environment, technically by learning an action-conditioned future representation/scene regressor. This allows the agent to learn action policies (i.e., visuomotor policies) by interacting with the dreaming model rather than the real-world. We experimentally confirm that our dreaming model enables robot learning of policies that transfer to the real-world.