Self-Supervised Exploration via Disagreement
This work addresses the challenge of scalable and robust exploration for robotics and stochastic domains, offering a novel approach that is more efficient than prior methods.
The paper tackles the problem of efficient exploration in stochastic environments and real robotics by proposing a self-supervised method that maximizes disagreement among an ensemble of dynamics models, enabling sample-efficient learning without external rewards. It demonstrates efficacy in benchmark environments like stochastic-Atari, Mujoco, and Unity, and successfully implements the method on a real robot learning from scratch.
Efficient exploration is a long-standing problem in sensorimotor learning. Major advances have been demonstrated in noise-free, non-stochastic domains such as video games and simulation. However, most of these formulations either get stuck in environments with stochastic dynamics or are too inefficient to be scalable to real robotics setups. In this paper, we propose a formulation for exploration inspired by the work in active learning literature. Specifically, we train an ensemble of dynamics models and incentivize the agent to explore such that the disagreement of those ensembles is maximized. This allows the agent to learn skills by exploring in a self-supervised manner without any external reward. Notably, we further leverage the disagreement objective to optimize the agent's policy in a differentiable manner, without using reinforcement learning, which results in a sample-efficient exploration. We demonstrate the efficacy of this formulation across a variety of benchmark environments including stochastic-Atari, Mujoco and Unity. Finally, we implement our differentiable exploration on a real robot which learns to interact with objects completely from scratch. Project videos and code are at https://pathak22.github.io/exploration-by-disagreement/