Automatically evaluating question-answering models
7%. That's the error rate of Amazon's new AI evaluation method compared to human reviewers.

Why it matters
This automation breakthrough could dramatically reduce costs and time for enterprises deploying Q&A systems, making AI quality assurance scalable for business applications.
The key facts
2 to know7% error rate compared to human evaluation
Method automatically evaluates question-answering models
Go to the source
Amazon Scienceamazon.science
Publisher excerpt: Relative to human evaluation of question-answering models, the new method has an error rate of only 7%.

