Dynamic model selection in speech-to-text processing

A system for converting spoken words into text first listens to the sounds and makes an initial guess about what was said, also noting how confident it is. If it's very confident, it picks that word. If it's unsure, it then consults a second system that understands language structure and context to help choose the correct word, specifically avoiding the second system if the first is confident to save computing power.

Why it matters: Filed when speech models were evolving, today's larger, more resource-intensive language models make this selective approach to saving processing power even more critical and efficient.

Status: Not in forceListed as no longer active. The specific reason is not in the record we hold.
How hard to build: SpecializedSoftware logic, acoustic and language models, training.

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