AnyMorph: Learning Transferable Polices By Inferring Agent Morphology
This addresses the challenge of eliminating re-training for new agents in robotics or simulation, though it is incremental as it builds on prior morphology-agnostic approaches.
The paper tackles the problem of training reinforcement learning policies that can transfer to new agent morphologies without requiring hand-designed descriptions, achieving state-of-the-art zero-shot generalization on a standard benchmark.
The prototypical approach to reinforcement learning involves training policies tailored to a particular agent from scratch for every new morphology. Recent work aims to eliminate the re-training of policies by investigating whether a morphology-agnostic policy, trained on a diverse set of agents with similar task objectives, can be transferred to new agents with unseen morphologies without re-training. This is a challenging problem that required previous approaches to use hand-designed descriptions of the new agent's morphology. Instead of hand-designing this description, we propose a data-driven method that learns a representation of morphology directly from the reinforcement learning objective. Ours is the first reinforcement learning algorithm that can train a policy to generalize to new agent morphologies without requiring a description of the agent's morphology in advance. We evaluate our approach on the standard benchmark for agent-agnostic control, and improve over the current state of the art in zero-shot generalization to new agents. Importantly, our method attains good performance without an explicit description of morphology.