/Patent Tracker
GooglePatent applicationAI & softwareInterest 6/10

Google patents secure on-device AI that adds hidden context to predictions

Patent drawing
Patent drawing from the official filing.

Google describes a system that runs AI models directly on your phone or computer while keeping sensitive data secure. It uses a protected area called an enclave to mix your app's data with hidden device details, like location or time, before the AI makes a guess. This allows apps to get smarter predictions without directly seeing the private information.

Why it mattersThis approach could make personal AI features more private by ensuring apps can't peek at the extra details used to improve their accuracy.

In the filing’s own words

The present disclosure provides systems and methods for on-device machine learning. In particular, the present disclosure is directed to an on-device machine learning platform and associated techniques that enable on-device prediction, training, example collection, and/or other machine learning tasks or functionality. The on-device machine learning platform can include a context provider that securely injects context features into collected training examples and/or client-provided input data used to generate predictions/inferences. Thus, the on-device machine learning platform can enable centralized training example collection, model training, and usage of machine-learned models as a service to applications or other clients.

What exactly they claim (claim 1)

1. A computer-implemented method comprising: accessing an enclave on a computing device for implementing operations of an on-device platform for implementation of machine-learned models on behalf of an application operating on the computing device; receiving, via an interface of the on-device platform, application data from the application; accessing, via a context provider, one or more context features descriptive of a context associated with the computing device; providing, by the on-device platform and using the enclave, the one or more context features for input into a machine-learned model alongside the application data in a model input, wherein the one or more context features are not directly accessible in the enclave to the application; obtaining, by the on-device platform, an inference generated by the machine-learned model based on the model input; and returning an output based on the inference to the application.

Facts

Company
Google (Google LLC)
Official title
ON-DEVICE MACHINE LEARNING PLATFORM
Publication number
US20260311067A1
Application number
19717174
Filed
Jun 23, 2026
Published
Oct 8, 2026
Inventors
5
Claims
20

Classified under

G06N 20/00 Physics & computingG06N 5/04 Physics & computing
Read the full filing (official PDF) ↗ Open in the Patent Tracker

More from Google in AI & software

Summaries are simplified from the filing’s own abstract and first claim. A patent filing is not a product announcement.
Patent data and drawings: U.S. Patent and Trademark Office (USPTO). Nebula is not affiliated with or endorsed by the USPTO.