Instant Policy: In-Context Imitation Learning via Graph Diffusion
This addresses the challenge of flexible and efficient robot task learning for robotics applications, representing a novel method rather than an incremental improvement.
The paper tackles the problem of enabling robots to learn new tasks instantly from one or two demonstrations without further training, achieving this through in-context imitation learning with a graph diffusion model, which shows rapid learning in simulated and real experiments.
Following the impressive capabilities of in-context learning with large transformers, In-Context Imitation Learning (ICIL) is a promising opportunity for robotics. We introduce Instant Policy, which learns new tasks instantly (without further training) from just one or two demonstrations, achieving ICIL through two key components. First, we introduce inductive biases through a graph representation and model ICIL as a graph generation problem with a learned diffusion process, enabling structured reasoning over demonstrations, observations, and actions. Second, we show that such a model can be trained using pseudo-demonstrations - arbitrary trajectories generated in simulation - as a virtually infinite pool of training data. Simulated and real experiments show that Instant Policy enables rapid learning of various everyday robot tasks. We also show how it can serve as a foundation for cross-embodiment and zero-shot transfer to language-defined tasks. Code and videos are available at https://www.robot-learning.uk/instant-policy.