Generating command-specific language model discourses for digital assistant interpretation

A digital assistant uses this system to better understand spoken commands. It works by creating a specialized language model from command templates, which are linked to specific actions. These templates are used to generate many example phrases, which are then tagged and organized into clusters. When a user speaks a command, the system matches it to these examples to determine the intended action. The claims specify that these example phrases are generated by searching external data and extracting relevant content.

Why it matters: Filed before the widespread adoption of large language models. Modern LLMs can now generate the diverse, high-quality synthetic documents described in the claims much more efficiently and effectively, simplifying a key part of the system.

Status: Not in forceListed as no longer active. The specific reason is not in the record we hold.
How hard to build: SpecializedComplex software system, language model, data processing

3 AI reinvention directions for this patent

See all three

AI gives you a few directions you could take this. Pick one, and we check whether your version is different enough to patent, then write the filing.

Reinvent this with AI

Where this one came from

Every morning we go through millions of abandoned patents and pick out 15 worth your time. This was one of today’s.

This one

+10