Aligning past experience with the robot's current physical reality (IMAGE)

Science China Press

Aligning past experience with the robot's current physical reality

Caption

OMPO uses a large global replay buffer of heterogeneous history and a small local buffer of recent interactions. Transition Occupancy Matching aligns historical data with the current environment across policy shifts, dynamics shifts and their combination, helping the agent learn stable behavior.

Credit

Yu Luo et al., National Science Review (CC BY 4.0)

Usage Restrictions

Credit must be given to the creator.

License

CC BY

Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.