Changeflow GovPing Telecom & Technology US12608585B2: GDM Holdings Neural Network Patent
Routine Rule Added Final

US12608585B2: GDM Holdings Neural Network Patent

Favicon for changeflow.com USPTO Patent Grants - AI & Computing (G06N)
Published
Detected
Email

Summary

USPTO granted patent US12608585B2 to GDM HOLDING LLC on April 21, 2026. The patent covers a neural network system for robotic agents, combining simulation-trained deep neural networks with robot-trained networks to process environmental observations and generate policy outputs defining robotic actions. Six inventors are named: Razvan Pascanu, Raia Thais Hadsell, Mel Vecerik, Thomas Rothoerl, Andrei-Alexandru Rusu, and Nicolas Manfred Otto Heess. The patent contains 22 claims across CPC classifications G06N 3/008, G06N 3/092, G06N 3/10, G06N 3/045, G06N 3/08, G06N 3/082, and G06N 20/00.

“The neural network system is configured to receive an observation characterizing a current state of a real-world environment being interacted with by a robotic agent to perform a robotic task and to process the observation to generate a policy output that defines an action to be performed by the robotic agent in response to the observation.”

USPTO , verbatim from source
Published by USPTO on changeflow.com . Detected, standardized, and enriched by GovPing. Review our methodology and editorial standards .

About this source

GovPing monitors USPTO Patent Grants - AI & Computing (G06N) for new telecom & technology regulatory changes. Every update since tracking began is archived, classified, and available as free RSS or email alerts — 32 changes logged to date.

What changed

USPTO issued patent US12608585B2 to GDM HOLDING LLC, covering a neural network system that processes observations of a robotic agent's real-world environment and generates policy outputs to define actions. The system combines a simulation-trained deep neural network (trained in a simulated environment) with a robot-trained DNN that maps observations directly to policy outputs. The patent names Razvan Pascanu, Raia Thais Hadsell, Mel Vecerik, Thomas Rothoerl, Andrei-Alexandru Rusu, and Nicolas Manfred Otto Heess as inventors. Twenty-two claims span neural network architectures (G06N 3/008, G06N 3/092, G06N 3/10, G06N 3/045, G06N 3/08, G06N 3/082) and machine learning applications (G06N 20/00). Competitors developing reinforcement learning systems for robotic control should review this intellectual property for freedom-to-operate implications.

Archived snapshot

Apr 23, 2026

GovPing captured this document from the original source. If the source has since changed or been removed, this is the text as it existed at that time.

← USPTO Patent Grants

Neural networks for selecting actions to be performed by a robotic agent

Grant US12608585B2 Kind: B2 Apr 21, 2026

Assignee

GDM HOLDING LLC

Inventors

Razvan Pascanu, Raia Thais Hadsell, Mel Vecerik, Thomas Rothoerl, Andrei-Alexandru Rusu, Nicolas Manfred Otto Heess

Abstract

A system includes a neural network system implemented by one or more computers. The neural network system is configured to receive an observation characterizing a current state of a real-world environment being interacted with by a robotic agent to perform a robotic task and to process the observation to generate a policy output that defines an action to be performed by the robotic agent in response to the observation. The neural network system includes: (i) a sequence of deep neural networks (DNNs), in which the sequence of DNNs includes a simulation-trained DNN that has been trained on interactions of a simulated version of the robotic agent with a simulated version of the real-world environment to perform a simulated version of the robotic task, and (ii) a first robot-trained DNN that is configured to receive the observation and to process the observation to generate the policy output.

CPC Classifications

G06N 3/008 G06N 3/092 G06N 3/10 G06N 3/045 G06N 3/08 G06N 3/082 G06N 20/00

Filing Date

2022-07-25

Application No.

17872528

Claims

22

View original document →

Get daily alerts for USPTO Patent Grants - AI & Computing (G06N)

Daily digest delivered to your inbox.

Free. Unsubscribe anytime.

About this page

What is GovPing?

Every important government, regulator, and court update from around the world. One place. Real-time. Free. Our mission

What's from the agency?

Source document text, dates, docket IDs, and authority are extracted directly from USPTO.

What's AI-generated?

The summary, classification, recommended actions, deadlines, and penalty information are AI-generated from the original text and may contain errors. Always verify against the source document.

Last updated

Classification

Agency
USPTO
Published
April 21st, 2026
Instrument
Rule
Branch
Executive
Legal weight
Binding
Stage
Final
Change scope
Minor

Who this affects

Applies to
Technology companies Manufacturers
Industry sector
5112 Software & Technology
Activity scope
Patent grant Neural network systems Robotic agents
Geographic scope
United States US

Taxonomy

Primary area
Intellectual Property
Operational domain
Legal

Get alerts for this source

We'll email you when USPTO Patent Grants - AI & Computing (G06N) publishes new changes.

Free. Unsubscribe anytime.

You're subscribed!