One-Shot Visual Imitation Learning via Meta-Learning

Abstract

In order for a robot to be a generalist that can perform a wide range of jobs, it must be able to acquire a wide variety of skills quickly and efficiently in complex unstructured environments. High-capacity models such as deep neural networks can enable a robot to represent complex skills, but learning each skill from scratch then becomes infeasible. In this work, we present a meta-imitation learning method that enables a robot to learn how to learn more efficiently, allowing it to acquire new skills from just a single demonstration. Unlike prior methods for one-shot imitation, our method can scale to raw pixel inputs and requires data from significantly fewer prior tasks for effective learning of new skills. Our experiments on both simulated and real robot platforms demonstrate the ability to learn new tasks, end-to-end, from a single visual demonstration.

Conference on Robot Learning, 2017 (to appear). First two authors contributed equally. Video available at https://sites.google.com/view/one-shot-imitation

References in corpus (1)

• Learning Invariant Feature Spaces to Transfer Skills with Reinforcement Learning

Cited by in corpus (5)

• MetaPred: Meta-Learning for Clinical Risk Prediction with Limited Patient Electronic Health Records
• PLOTS: Procedure Learning from Observations using Subtask Structure
• Knowledge Transfer Between Robots with Similar Dynamics for High-Accuracy Impromptu Trajectory Tracking
• Extending Policy from One-Shot Learning through Coaching
• On the potential for open-endedness in neural networks