Get the newsletter
Nvidia · AI / ML · Filed Feb 12, 2026 · USPTO published Aug 13, 2026 · Aug 13 release

Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning

Nvidia has filed a patent for a system that figures out which neural network design works best for a task, without any of the organizations involved ever having to share their raw data with each other.

An autonomous vehicle fitted with sensors, cameras, and lidar units for environmental sensing and navigation. Drawing from patent filing US 2026/0236742 A1.
An autonomous vehicle fitted with sensors, cameras, and lidar units for environmental sensing and navigation.All 58 drawings ↓
Publication number
US 2026/0236742 A1(US20260236742A1)
Applicant
NVIDIA Corporation
Filed · Published
Feb 12, 2026 · Aug 13, 2026
Inventors
Holger Reinhard Roth, Dong Yang, Wenqi Li, Andriy Myronenko, Wentao Zhu, Ziyue Xu, Xiaosong Wang, Daguang Xu
Status when we published
Waiting for an examiner
Document
20 claims
US classification
706/15
Parent application
Continuation of 16/889,652 (filed Jun 1, 2020)(16889652)

How Nvidia trains AI across private data silos

Ever wondered how hospitals could contribute to a shared AI model without sending patient records outside their walls? That's exactly the problem this patent is trying to solve, and it turns out the usual answer, gathering all the data in one place and training there, simply isn't allowed in most regulated industries.

Nvidia's approach is to let each organization train its piece of the AI model on its own data, then share only the learned patterns (not the underlying records) back to a central coordinator. The coordinator uses those contributions to figure out not just how well the model is learning, but which model design works best across all the different data sets involved.

The clever part is that different organizations may be using different types of data, and the system can match which portions of the AI network are relevant for which data type. You end up with a better-tuned model without anyone ever handing over their private information.

From the filing · CLAIM 1
… cause one or more parameters of different portions of one or more first neural networks to be updated using one or more parameters of correspondingly different portions of two or more second neural networks, wherein the different portions of the two or more second neural networks correspond to the type of training data …

Translation: The system updates parts of a neural network by pulling in specific data patterns from other networks trained on similar info.

How the federated architecture search actually selects a model

The patent describes a process called federated neural architecture search, which combines two existing ideas in a single workflow.

The first idea is federated learning (training a shared AI model across multiple independent computers or institutions, where only model updates travel over the network, never the raw data). The second is neural architecture search (automatically testing many possible AI network designs to find the one that performs best on a given task, rather than having engineers pick the structure by hand).

The system works roughly like this:

  • Each participating organization (a hospital, a research lab, a company) trains a local version of the AI on its own private data.
  • The type of data each site holds is noted, and only the portions of the model that are relevant to that data type are updated from that site's results.
  • A coordinating server collects those partial updates and uses them to compare candidate network designs, selecting whichever architecture performs best across all the distributed sites.

The claim is careful to say the selection is "based, at least in part, on the type of training data" at each site, which means the architecture search is data-aware: a design that works well for radiology scans at one hospital can be weighted differently than one tuned for electronic health records at another.

From the filing · THE ABSTRACT
Apparatuses, systems, and techniques to select a neural network architecture from a plurality of neural networks in a federated learning (FL) setting. In at least one embodiment, a neural network is trained by combining training results from different FL computing systems …

Translation: This technology picks the best AI structure by merging results from multiple computers that learn without sharing raw data.

What this means for AI trained on medical or sensitive data

Healthcare AI is the most obvious application. Training diagnostic models today usually requires either moving patient data to a central server (legally complicated and politically sensitive) or accepting that the model only sees one institution's patient population (statistically limiting). A system that can search for the best model design while respecting each institution's data boundaries could make AI diagnostics both more accurate and more broadly deployable.

The broader implication is that federated architecture search could apply anywhere data is fragmented and protected, from financial fraud detection across competing banks to predictive maintenance across manufacturing rivals. Nvidia already sells the hardware and software stacks (CUDA, Clara, Flare) that underpin these workflows, so a patent in this space fits a clear product strategy. For anyone tracking how AI infrastructure is maturing under real-world privacy constraints, this patent sits alongside a steady stream of new Big Tech patents pushing federated and privacy-preserving AI methods into specialized, regulated industries.

Source: US 2026/0236742 A1 at the USPTO (PDF). Full patent text and figures are in the official publication.

Get our take in your Top Stories

Liked this breakdown? Add Patentlyze as a preferred source on Google, and our plain-English take shows up more often in your Top Stories the next time Nvidia patent news breaks.

There are more where this came from

We read every patent application Big Tech publishes and send you the ones worth knowing. Plain English, free, every week.

The drawings · 58 sheets

58 drawing sheets from US 2026/0236742 A1 · click any drawing to enlarge

Federated neural network client server system, drawing sheet 1 from US 2026/0236742 A1
Sheet 1 Federated neural network client server system
Superblock candidate architecture for image processing, drawing sheet 2 from US 2026/0236742 A1
Sheet 2 Superblock candidate architecture for image processing
Convolutional neural network encoder decoder pipeline, drawing sheet 3 from US 2026/0236742 A1
Sheet 3 Convolutional neural network encoder decoder pipeline
Dynamic path selection across neural network candidates, drawing sheet 4 from US 2026/0236742 A1
Sheet 4 Dynamic path selection across neural network candidates
Flowchart of federated supernet training steps, drawing sheet 5 from US 2026/0236742 A1
Sheet 5 Flowchart of federated supernet training steps
Flowchart of data inference and network selection, drawing sheet 6 from US 2026/0236742 A1
Sheet 6 Flowchart of data inference and network selection
Flowchart of data processing with varying architectures, drawing sheet 7 from US 2026/0236742 A1
Sheet 7 Flowchart of data processing with varying architectures
Performance comparison box plots across datasets, drawing sheet 8 from US 2026/0236742 A1
Sheet 8 Performance comparison box plots across datasets
Show the other 50 sheets
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 9
Sheet 9
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 10
Sheet 10
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 11
Sheet 11
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 12
Sheet 12
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 13
Sheet 13
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 14
Sheet 14
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 15
Sheet 15
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 16
Sheet 16
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 17
Sheet 17
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 18
Sheet 18
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 19
Sheet 19
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 20
Sheet 20
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 21
Sheet 21
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 22
Sheet 22
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 23
Sheet 23
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 24
Sheet 24
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 25
Sheet 25
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 26
Sheet 26
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 27
Sheet 27
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 28
Sheet 28
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 29
Sheet 29
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 30
Sheet 30
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 31
Sheet 31
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 32
Sheet 32
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 33
Sheet 33
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 34
Sheet 34
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 35
Sheet 35
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 36
Sheet 36
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 37
Sheet 37
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 38
Sheet 38
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 39
Sheet 39
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 40
Sheet 40
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 41
Sheet 41
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 42
Sheet 42
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 43
Sheet 43
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 44
Sheet 44
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 45
Sheet 45
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 46
Sheet 46
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 47
Sheet 47
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 48
Sheet 48
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 49
Sheet 49
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 50
Sheet 50
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 51
Sheet 51
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 52
Sheet 52
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 53
Sheet 53
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 54
Sheet 54
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 55
Sheet 55
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 56
Sheet 56
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 57
Sheet 57
Nvidia Patent Selects the Optimal Neural Network Architecture Through Federated Learning patent drawing, sheet 58
Sheet 58
Patent filing page

More in AI training & infrastructure

All 316 AI training & infrastructure patents →

Watchlists this filing belongs to

All watchlists →