Training AI Models Across Different Computers
This 2026 patent describes a way to train AI models on one computer, send a version to another computer for further training with private data, and then update the original model with the improvements.
Original patent title: “Distributed deep learning using a distributed deep neural network”
This 2026 patent describes a way to train AI models on one computer, send a version to another computer for further training with private data, and then update the original model with the improvements. Granted to Deep Sentinel in 2026 with 21 claims, and it is expected to expire in 2037.
Key facts
Coverage
What does this patent actually cover?
This patent details a method for distributed AI training. A central computer (first host system) trains an initial neural network using data that has already been filtered from multiple remote sources. This initial network is then sent to a remote computer (second host system). The remote computer further trains this network using its own private data, creating a customized version. This customized network is then installed and used on the remote computer to process its live data stream. Crucially, the remote computer sends updated coefficients back to the central computer, allowing the original network to be improved based on the private data insights gained remotely. For example, a central server could train a general security camera AI, send it to individual homes, where each home's camera further trains it on local activity, and then sends back anonymized updates to improve the central AI.
The gap
What does this patent NOT cover?
- Training that only occurs on a single computer system.
- Using a neural network that is not further trained on private, local data at the second host system.
- Sending raw, unfiltered event data from the remote systems to the first host system.
- Updating the central neural network without receiving updated coefficients from a remote system.
- Training that does not involve evaluating the neural network at both the first and second host systems.
These exclusions are unique to PatentBrief — derived from the actual claim language, not patent-office boilerplate.
What made this novel
The innovation lies in the two-way learning process: a central model is improved by private, local data without that data ever leaving the local system, and the central model, in turn, helps filter data for the local system. This creates a continuous, privacy-preserving feedback loop for AI model improvement.
The Patent Drawing
Schematic visualization of the patent's claim structure. Hand-drawn diagrams in progress for each landmark patent.
Where you've seen this
Real-world examples
Security camera systems that learn local activity patterns.
Smart home devices that personalize their functions based on user behavior.
Industrial monitoring systems that adapt to specific machine behaviors.
Why it matters
The bigger picture
This patent addresses the challenge of training powerful AI models while respecting data privacy. It enables AI systems, like those used in security or smart home devices, to learn from diverse, real-world data without needing to collect all that sensitive data in one central location. This approach is key for developing intelligent systems that can adapt to local conditions while maintaining a globally improved model.
Filed
April 19, 2017
Granted
March 10, 2026
Market context
Who's building on this
Companies in this space
Deep Sentinel Corp, the assigneeassigneeThe entity that owns the patent — usually the inventor's employer or a company.Read more →, is likely developing systems based on this patent for their own security products. The principles of distributed and federated learning are also foundational to many modern AI platforms, including those from major tech companies like Google (Federated Learning) and Apple, who are continuously refining methods for training models on device.
Market impact
This patent, granted in 2026, describes a method that has become increasingly important for privacy-preserving AI. It enables the development of intelligent edge devices that can learn and adapt locally, reducing reliance on massive cloud data aggregation and addressing growing privacy concerns. This approach is critical for the expansion of AI into consumer electronics and IoT devices.
Claim 1 — Plain English
What this patent covers
This patent details a method for distributed AI training. A central computer (first host system) trains an initial neural network using data that has already been filtered from multiple remote sources. This initial network is then sent to a remote computer (second host system). The remote computer further trains this network using its own private data, creating a customized version. This customized network is then installed and used on the remote computer to process its live data stream. Crucially, the remote computer sends updated coefficients back to the central computer, allowing the original network to be improved based on the private data insights gained remotely. For example, a central server could train a general security camera AI, send it to individual homes, where each home's camera further trains it on local activity, and then sends back anonymized updates to improve the central AI.
The clever bit
The innovation lies in the two-way learning process: a central model is improved by private, local data without that data ever leaving the local system, and the central model, in turn, helps filter data for the local system. This creates a continuous, privacy-preserving feedback loop for AI model improvement.
What it does not cover
- Training that only occurs on a single computer system.
- Using a neural network that is not further trained on private, local data at the second host system.
- Sending raw, unfiltered event data from the remote systems to the first host system.
- Updating the central neural network without receiving updated coefficients from a remote system.
- Training that does not involve evaluating the neural network at both the first and second host systems.
Patent timeline
Application submitted to the patent office
Application published, typically 18 months after filing
Patent officially issued
Patent enters public domain
PatentBrief Score
Impact Score
Early stage
Citation count
0/40
No citations yet
Claim breadth
14/20
Broad claimsclaimsThe numbered statements at the end of a patent that legally define what the inventor owns.Read more →
Recency
20/20
Granted within 5 years
Assignee scale
0/20
Independent or smaller assigneeassigneeThe entity that owns the patent — usually the inventor's employer or a company.Read more →
PatentBrief Impact Score — based on citation count, claim breadth, recency, and assignee scale. Not a legal assessment.
Heuristic Value Estimate
What this patent might be worth
$31K – $100K
Midpoint $62K · 10.8 yr remaining · industry ×1.6
Heuristic only — blends forward/backward citation counts, claim scope, time remaining, litigation history, and CPC-derived industry baseline. Real valuations need a professional appraisal.
The original legal language
Original claims
21 claims as filed with the patent office.
Concepts involved
Citations
Patent lineage
Cite this patent
Chen, C., Yip, C., & Selinger, D. L. (2026). Training AI Models Across Different Computers (U.S. Patent No. 12,574,477). U.S. Patent and Trademark Office. https://patentbrief.org/patent/us/12574477/distributed-deep-learning-using-a-distributed-deep-neural-network
Auto-generated from the patent record. Double-check author order and the issue date against the official USPTO document before submitting.
Embed
Add this patent to your site
Drop this plain-English patent card into any blog post or article — free, no signup. It always links back to the full breakdown here.
<div data-patentlens-widget data-patent-number="US12574477"></div> <script src="https://patentbrief.org/embed.js" async></script>
Stay in the loop
Get a weekly digest of new patents.
One email per week. No spam. Unsubscribe anytime.
Keep exploring
Related patents you should know
US 4683195 · 1987
How to Make Billions of Copies of a DNA Segment
This patent describes the Polymerase Chain Reaction (PCR), a method to rapidly create many copies of a specific piece of DNA or RNA, enabling its detection and analysis.
Cetus Corp
US 8697359 · 2014
How to Edit Genes in Human Cells Using an Engineered CRISPR System
This patent describes an engineered CRISPR-Cas9 system for precisely cutting DNA in eukaryotic cells to change how genes work, opening the door for gene editing in complex organisms.
Massachusetts Institute of Technology
US 7657849 · 2010
How the iPhone's Slide-to-Unlock Gesture Works
Apple's 2010 patent describes unlocking a device by dragging a specific graphical image across the touchscreen along a predefined path, a gesture that became iconic with the original iPhone.
Apple Inc
US 4733665 · 1988
How Doctors Implant a Permanent Stent Using a Balloon
This patent describes the method for placing a permanent, expandable wire mesh tube inside a blood vessel or other body tube using a balloon-tipped catheter to widen it and keep it open.
Expandable Grafts Partnership
US 4965188 · 1990
How to Make Many Copies of a DNA Piece with Heat
This patent describes the Polymerase Chain Reaction (PCR) method, a technique to make millions of copies of a specific DNA segment using a heat-resistant enzyme and repeated temperature changes.
Cetus Corp
US 4235871 · 1980
How to Encapsulate Active Materials in Lipid Bubbles Efficiently
This patent describes a method for trapping biologically active substances inside tiny, multi-layered fat bubbles called liposomes, using a specific water-in-oil emulsion and gel-forming process to improve how much material gets captured.
Individual
Semantically similar
You might also find these interesting
US 12518214 · 2026 · Nant Holdings IP
Training AI on Private Data Without Seeing It
US 12423586 · 2025 · D5AI
Making AI Smarter by Focusing on Unsure 'Nodes'
US 20230162023 · Mitsubishi Electric Research Laboratories
Automated AI for Adapting to New Data Without Retraining
US 10410117 · 2019 · BrainChip Inc
How to Save and Reuse Skills Learned by Artificial Intelligence Hardware
More to explore
More in Consumer Electronics
US 7657849 · 2010 · Apple Inc
How the iPhone's Slide-to-Unlock Gesture Works
US 7479949 · 2009 · Apple Inc
How Touchscreens Understand Your Finger Swipes and Scrolls
US 4528643 · 1985 · FPDC Inc
How Stores Make Custom Products On-Demand with Remote Approval
US 7469381 · 2008 · Apple Inc
How Touchscreens Show and Snap Back When You Scroll Past an Edge
New to patents?
Common Questions
Frequently Asked Questions
What does Training AI Models Across Different Computers cover?
This 2026 patent describes a way to train AI models on one computer, send a version to another computer for further training with private data, and then update the original model with the improvements.
Who owns patent US 12574477?
Deep Sentinel owns this patent, granted in 2026.
When does this patent expire?
This patent is expected to expire on April 19, 2037, when the invention enters the public domain.
What problem does this patent solve?
This patent addresses the challenge of training powerful AI models while respecting data privacy. It enables AI systems, like those used in security or smart home devices, to learn from diverse, real-world data without needing to collect all that sensitive data in one central location. This approach is key for developing intelligent systems that can adapt to local conditions while maintaining a globally improved model.
What does this patent NOT cover?
Training that only occurs on a single computer system.
Patent monitoring