Label confidence scoring

This invention describes a system that evaluates how good a human's transcription of spoken words is. It takes human speech and the human's written version of it, then uses a computer model to calculate a 'confidence score' for that transcription, possibly considering details like when or where the transcription was made. The system can even ask the human for their own confidence rating to help improve its scoring.

Why it matters: Filed before AI speech-to-text models achieved near-human accuracy. The invention's method for scoring human transcription quality is now highly relevant for evaluating human corrections of AI-generated text, or for training AI models to self-assess their confidence.

Status: Not in forceListed as no longer active.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

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