How to Update AI on Small Devices with Slow Internet
This patent describes a method for efficiently updating artificial intelligence models on small, internet-connected devices, like smart cameras, by sending only the changes, or 'patches,' instead of the entire updated model, which saves bandwidth.
Original patent title: “Systems and Methods for Deploying and Updating Neural Networks at the Edge of a Network”
This patent describes a method for efficiently updating artificial intelligence models on small, internet-connected devices, like smart cameras, by sending only the changes, or 'patches,' instead of the entire updated model, which saves bandwidth. Owned by Ubotica Technologies with 18 claims, and it is expected to expire in 2045.
Coverage
What does this patent actually cover?
This patent describes a system for keeping neural networks on 'edge devices' up-to-date, especially when those devices have slow internet connections. First, a powerful central computer trains a neural network and sends this initial version to the edge device (ClaimclaimA numbered sentence at the end of a patent that legally defines what the inventor owns. The most important section.Read more → 1). The edge device then collects data, uses the neural network, and sends back specific pieces of information, such as parts of the data it collected, internal calculations called 'activations,' or its final 'inference results' (Claim 4). The central computer uses this feedback to create an improved version of the neural network. Instead of sending the entire updated network back, the central computer generates a 'neural network difference model' (a patch) by comparing the new and old networks (Claim 1). This patch identifies only the changes, using techniques like 'layer freezing' or 'weights freezing' based on specific criteria, and is then sent to the edge device (Claim 1). For example, a smart security camera with a slow cellular connection could receive small updates to its object recognition AI without needing to download a massive new software package each time.
The gap
What does this patent NOT cover?
- Does not cover sending the entire updated neural network to the edge device, as it specifically focuses on sending a 'neural network difference model' (patch).
- Does not cover updating neural networks on edge devices without a centralized site/device performing the initial training and update generation.
- Does not cover generating the neural network difference model using techniques other than 'layer freezing' or 'weights freezing' with 'minimum size' or 'minimum delta' methods.
- Does not cover edge devices that have high-bandwidth uplink capability, as the claimsclaimsThe numbered statements at the end of a patent that legally define what the inventor owns.Read more → specifically address 'low-bandwidth uplink capability'.
- Does not cover scenarios where the edge device sends back information other than portions of a dataset, activations, or overall inference results.
These exclusions are unique to PatentBrief — derived from the actual claim language, not patent-office boilerplate.
Key facts
What made this novel
The truly clever part is sending only the 'difference' or 'patch' of the neural network rather than the entire updated model. This significantly reduces the amount of data that needs to be sent to edge devices, especially those with slow internet, by intelligently identifying and packaging only the changed parts of the AI model.
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
AI-powered security cameras in remote locations
Smart factory sensors monitoring equipment health
Agricultural robots performing crop analysis
Autonomous vehicles receiving targeted AI updates
Smart home devices with local AI processing
Why it matters
The bigger picture
As more artificial intelligence moves from cloud data centers to local devices like smart home gadgets and industrial sensors, efficiently updating these AI models becomes crucial. This patent addresses a key challenge for 'edge AI' by enabling updates over limited internet connections. It could help ensure that devices deployed in remote areas or with constrained network access can still benefit from continuous AI improvements. This method helps reduce data transfer costs and speeds up the deployment of new AI capabilities to a vast number of devices.
Filed
August 7, 2025
Market context
Who's building on this
Companies in this space
Companies developing edge AI solutions are actively working on efficient model deployment and update strategies. Major cloud providers like Amazon Web Services (AWS) with AWS IoT Greengrass, Microsoft with Azure IoT Edge, and Google Cloud with Edge TPU are building platforms to manage and update AI on edge devices. Additionally, specialized AI hardware and software startups focus on optimizing neural network performance and updates for constrained environments.
Market impact
This patent addresses a growing need in the 'edge AI' market, where the ability to efficiently update AI models on deployed devices is critical for long-term functionality and performance. If widely adopted, such methods could reduce operational costs for companies managing large fleets of AI-enabled edge devices by minimizing data transfer. It could also enable more frequent and granular AI improvements, leading to better performance and new capabilities for products in industries ranging from smart cities to industrial automation.
Claim 1 — Plain English
What this patent covers
This patent describes a system for keeping neural networks on 'edge devices' up-to-date, especially when those devices have slow internet connections. First, a powerful central computer trains a neural network and sends this initial version to the edge device (Claim 1). The edge device then collects data, uses the neural network, and sends back specific pieces of information, such as parts of the data it collected, internal calculations called 'activations,' or its final 'inference results' (Claim 4). The central computer uses this feedback to create an improved version of the neural network. Instead of sending the entire updated network back, the central computer generates a 'neural network difference model' (a patch) by comparing the new and old networks (Claim 1). This patch identifies only the changes, using techniques like 'layer freezing' or 'weights freezing' based on specific criteria, and is then sent to the edge device (Claim 1). For example, a smart security camera with a slow cellular connection could receive small updates to its object recognition AI without needing to download a massive new software package each time.
The clever bit
The truly clever part is sending only the 'difference' or 'patch' of the neural network rather than the entire updated model. This significantly reduces the amount of data that needs to be sent to edge devices, especially those with slow internet, by intelligently identifying and packaging only the changed parts of the AI model.
What it does not cover
- Does not cover sending the entire updated neural network to the edge device, as it specifically focuses on sending a 'neural network difference model' (patch).
- Does not cover updating neural networks on edge devices without a centralized site/device performing the initial training and update generation.
- Does not cover generating the neural network difference model using techniques other than 'layer freezing' or 'weights freezing' with 'minimum size' or 'minimum delta' methods.
- Does not cover edge devices that have high-bandwidth uplink capability, as the claims specifically address 'low-bandwidth uplink capability'.
- Does not cover scenarios where the edge device sends back information other than portions of a dataset, activations, or overall inference results.
Patent timeline
Application submitted to the patent office
Patent enters public domain
PatentBrief Score
Impact Score
Limited data
Citation count
0/40
No citations yet
Claim breadth
12/20
Broad claimsclaimsThe numbered statements at the end of a patent that legally define what the inventor owns.Read more →
Recency
0/20
Older than 20 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 · 19.1 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
18 claims as filed with the patent office.
Concepts involved
Cite this patent
Dunne, A., & Buckley, F. How to Update AI on Small Devices with Slow Internet (U.S. Patent No. 20,250,363,357). U.S. Patent and Trademark Office. https://patentbrief.org/patent/us/20250363357/systems-and-methods-for-deploying-and-updating-neural-networks-at-the-edge-of-a-
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="US20250363357"></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 12574477 · 2026 · Deep Sentinel
Training AI Models Across Different Computers
US 20220383078 · Huawei Technologies Co
Adapting AI Models to Fit Device Resources
US 10832138 · 2020 · Samsung Electronics Co Ltd
How to Automatically Expand Neural Networks by Adding New Nodes
US 12443890 · 2025 · Google
How Devices Train Shared AI Models While Keeping Your Data Private
More to explore
More in AI & Machine Learning
US 10452978 · 2019 · Google LLC
How AI Models Understand Language Using 'Attention'
US 6523026 · 2003 · Huntsman International LLC
How Computers Find Hidden Connections Between Different Fields of Knowledge
US 11615208 · 2023 · Capital One Services LLC
How Cloud Systems Automatically Create and Train AI Data Models
US 10402750 · 2019 · Facebook Inc
How Facebook Uses Deep Learning to Predict What You Might Like
New to patents?
Common Questions
Frequently Asked Questions
What does How to Update AI on Small Devices with Slow Internet cover?
This patent describes a method for efficiently updating artificial intelligence models on small, internet-connected devices, like smart cameras, by sending only the changes, or 'patches,' instead of the entire updated model, which saves bandwidth.
Who owns patent US 20250363357?
This patent is owned by Ubotica Technologies.
When does this patent expire?
This patent is expected to expire on August 7, 2045, when the invention enters the public domain.
What problem does this patent solve?
As more artificial intelligence moves from cloud data centers to local devices like smart home gadgets and industrial sensors, efficiently updating these AI models becomes crucial. This patent addresses a key challenge for 'edge AI' by enabling updates over limited internet connections. It could help ensure that devices deployed in remote areas or with constrained network access can still benefit from continuous AI improvements. This method helps reduce data transfer costs and speeds up the deployment of new AI capabilities to a vast number of devices.
What does this patent NOT cover?
Does not cover sending the entire updated neural network to the edge device, as it specifically focuses on sending a 'neural network difference model' (patch).
Patent monitoring