Dynamics-Aware Unsupervised Discovery of Skills
This addresses the problem of poor generalization and complexity in MBRL for robotics and AI systems, offering a novel hybrid approach that is incremental but shows strong specific gains.
The paper tackles the difficulty of learning accurate global dynamics models in model-based reinforcement learning (MBRL) by proposing an unsupervised algorithm, Dynamics-Aware Discovery of Skills (DADS), that discovers predictable skills and learns their dynamics simultaneously. It demonstrates that zero-shot planning in the learned latent space outperforms standard MBRL and model-free goal-conditioned RL, handles sparse-reward tasks, and improves over prior hierarchical RL methods.
Conventionally, model-based reinforcement learning (MBRL) aims to learn a global model for the dynamics of the environment. A good model can potentially enable planning algorithms to generate a large variety of behaviors and solve diverse tasks. However, learning an accurate model for complex dynamical systems is difficult, and even then, the model might not generalize well outside the distribution of states on which it was trained. In this work, we combine model-based learning with model-free learning of primitives that make model-based planning easy. To that end, we aim to answer the question: how can we discover skills whose outcomes are easy to predict? We propose an unsupervised learning algorithm, Dynamics-Aware Discovery of Skills (DADS), which simultaneously discovers predictable behaviors and learns their dynamics. Our method can leverage continuous skill spaces, theoretically, allowing us to learn infinitely many behaviors even for high-dimensional state-spaces. We demonstrate that zero-shot planning in the learned latent space significantly outperforms standard MBRL and model-free goal-conditioned RL, can handle sparse-reward tasks, and substantially improves over prior hierarchical RL methods for unsupervised skill discovery.