OpenTechnical
This is a record of a document published elsewhere, first published by the United States Patent and Trademark Office, as US-2026077490-A1, on 19 March 2026. OpenTechnical did not first disclose it. The identifier, hashes and timestamps below record what this library served and when, not when the material became public. View the original.

OT-2026-000001734

Data-efficient hierarchical reinforcement learning

Published
Permanent URL
Classification
Robotics & Automation › Robot Architectures
Inventors
GU SHIXIANG, LEE HONGLAK, Levine Sergey
Assignee
GOOGLE LLC
Original source
US-2026077490-A1, the United States Patent and Trademark Office, 19 March 2026
Document SHA-256
d191671eb88560edc732bd27955db926018627e76f80e5539490948705fea197

Abstract

Training and/or utilizing a hierarchical reinforcement learning (HRL) model for robotic control. The HRL model can include at least a higher-level policy model and a lower-level policy model. Some implementations relate to technique(s) that enable more efficient off-policy training to be utilized in training of the higher-level policy model and/or the lower-level policy model. Some of those implementations utilize off-policy correction, which re-labels higher-level actions of experience data, generated in the past utilizing a previously trained version of the HRL model, with modified higher-level actions. The modified higher-level actions are then utilized to off-policy train the higher-level policy model. This can enable effective off-policy training despite the lower-level policy model being a different version at training time (relative to the version when the experience data was collected).

Publication information

Plain textMarkdownManifestIntegrity record

Citation

OpenTechnical Publication OT-2026-000001734, “Data-efficient hierarchical reinforcement learning,” published September 15, 2026 at 19:08:03 UTC, https://opentechnical.org/p/OT-2026-000001734.

To cite specific content, add the paragraph or figure numbers.

Terms

This record reproduces a document published elsewhere and is provided for reference and study. Rights in the original remain with its owner. See the publication policy.