/Patent Tracker
GoogleGranted patentAI & softwareInterest 6/10

Google patents method to train AI models using synthetic data

Google has patented a system that improves how AI models follow instructions by creating custom practice scenarios. The system uses one AI to generate specific tasks and answers, which are then used to adjust another AI model to better handle those specific jobs.

Why it mattersThis technique could help developers fine-tune AI systems for specific tasks without needing large amounts of real-world data, potentially making specialized AI tools more accurate.

In the filing’s own words

Aspects of the disclosed technology include computer-implemented systems and methods for adapting machine-learned models using high-quality synthetic data that is tailored to elicit improved instruction-following abilities for particular target instruction distributions and models. A model adaptation system can obtain instruction metadata indicative of at least one use case and at least one skill associated with a particular computing task to be performed by a target machine-learned model. The system can generate a metadata-conditioned synthetic instruction by prompting a generative model system including one or more machine-learned generative models with the instruction metadata as one or more constraints. The system can generate a model response by prompting the generative model system with the metadata-conditioned synthetic instruction. The system can modify a target sequence processing model based at least in part on a data pair including the metadata-conditioned synthetic instruction and the model response.

What exactly they claim (claim 1)

1. A computer-implemented method performed by one or more processors, the method comprising: obtaining instruction metadata indicative of at least one use case and at least one skill associated with a particular computing task to be performed by a target sequence processing model; generating a metadata-conditioned synthetic instruction by prompting a generative model system including one or more machine-learned generative models with the instruction metadata as one or more constraints; generating a model response by prompting the generative model system with the metadata-conditioned synthetic instruction; and modifying the target sequence processing model based at least in part on a data pair including the metadata-conditioned synthetic instruction and the model response.

Facts

Company
Google (Google LLC)
Official title
Machine-learned model alignment with synthetic data
Patent number
US12769552B2
Application number
19055348
Filed
Feb 17, 2025
Granted
Oct 20, 2026
Inventors
8
Claims
19

Classified under

G06F 9/30192 Physics & computingG06N 20/00 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.