Efficient adaption of robot control policy for new task using meta-learning based on meta-imitation learning and meta-reinforcement learning

US12083678B2 · kind B2 · utility

Assignee

• Google LLC · US

Inventors

• Mrinal Kalakrishnan · Palo Alto, US
• Yunfei Bai · Fremont, US
• Paul Wohlhart · Sunnyvale, US
• Eric Jang · Cupertino, US
• Chelsea Finn · Berkeley, US
• Seyed Mohammad Khansari Zadeh · San Carlos, US
• Sergey Levine · Redmond, US
• Allan Zhou · Mountain View, US
• Alexander Herzog · San Jose, US
• Daniel Kappler · San Francisco, US

Key dates

• Filing date: Jan 23, 2020
• Grant date: Sep 10, 2024
• Expiry date: Jan 17, 2041

Classification

• Technology area (CPC G): Physics
• CPC primary: G05B2219/40499
• WIPO field: Handling
• WIPO sector: Mechanical engineering

Abstract

Techniques are disclosed that enable training a meta-learning model, for use in causing a robot to perform a task, using imitation learning as well as reinforcement learning. Some implementations relate to training the meta-learning model using imitation learning based on one or more human guided demonstrations of the task. Additional or alternative implementations relate to training the meta-learning model using reinforcement learning based on trials of the robot attempting to perform the task. Further implementations relate to using the trained meta-learning model to few shot (or one shot) learn a new task based on a human guided demonstration of the new task.