/Patent Tracker
GooglePatent applicationAI & softwareInterest 6/10

Google patents system to help AI pick the right tools for tasks

Google has filed a patent for a framework that helps language models decide which specific tools or skills to use when answering questions. Instead of relying solely on its internal memory, which can lead to incorrect answers, the system matches the user's question to similar past examples to guide the AI toward the correct action.

Why it mattersThis could make AI assistants more reliable by reducing errors that happen when the model guesses instead of using a specific, verified tool.

In the filing’s own words

Provided are computer systems that enable machine-learned language models to choose the correct structural tools to leverage when solving challenging tasks. In particular, current language models rely heavily on internal knowledge to solve all downstream tasks, which causes hallucination and ungrounded answers. In contrast, proposed herein is a new tool-using framework for language models, which enables the language models to smartly route a query to the most relevant pre-defined skills.

What exactly they claim (claim 1)

1. A computer-implemented method for automating tool use based on learned similarity, the method comprising: obtaining, by a computing system comprising one or more computing devices, data descriptive of a query; generating, by the computing system, a query embedding for the query, wherein the query embedding is expressed within a latent embedding space; performing, by the computing system, a similarity search for the query embedding within the latent embedding space to identify one or more previously-defined embeddings associated with one or more previously-defined query-response pairs; generating, by the computing system, a prompt based on the query and the one or more previously-defined query-response pairs; providing, by the computing system, the prompt as an input for processing by a machine-learned language model; and receiving, by the computing system, a model response to the query that was output by the machine-learned language model based on processing of the prompt.

Facts

Company
Google (Google LLC)
Official title
Efficient use of Tools by Language Models
Publication number
US20260277967A1
Application number
19471305
Filed
Apr 1, 2024
Published
Sep 17, 2026
Inventors
5
Claims
20

Classified under

G06F 16/3344 Physics & computingG06F 16/35 Physics & computingG06F 40/284 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.