Amazon’s Augmented Artificial Intelligence (A21) platform is now generally available. It is a fully managed service designed to make it easy for developers to add human review to machine learning predictions so that they can review low confidence predictions made by AI.
According to Amazon, the solution can be used to help the machine extract financial information from scanned mortgage documents or an application that uses image recognition to identify counterfeit items online, so that the model becomes better over time.
RELATED CONTENT: Augmented intelligence will help, not replace, human workers
“This interplay between machine learning and human reviewers is critical to the success of machine learning systems, but human reviews are challenging and expensive to build and operate at scale, often involving multiple workflow steps, operating custom software to manage human review tasks and results, and recruiting and managing large groups of reviewers,” Amazon wrote in its announcement.
A21 aims to alleviate the challenging aspects of introducing human review with over 60 pre-built human review workflows for common machine learning to improve the symbiotic work with Amazon Rekognition and Amazon Textract, the company stated. Additionally, users only pay for each review that is needed rather than upfront.
Machine learning already has high confidence projections in fields such as identifying objects in images, extracting text from scanned documents, or transcribing and understanding spoken language, Amazon explained. However, in cases such as public safety and law enforcement, human review is strongly encouraged despite high confidence projections created by machine learning.
Developers can set up machine learning models in Amazon SageMaker through the Augmented AI console or via its Application Programming Interface (API). Then they can set a threshold for when results below a certain confidence score will be bounced to human reviewers.
Developers can specify the number of workers per review and Amazon A2I then routes each review to the precise number of reviewers.
Finally, human-validated results are stored in Amazon Simple Storage Service (S3), and developers can set up Amazon CloudWatch Events notifications to review metadata about inference accuracy and retrieve the results.
“Today, we’re excited to help our customers remove another obstacle to building machine learning applications with the launch of Amazon A2I, which makes it significantly easier and faster to incorporate human judgment into machine learning applications in order to ensure higher quality predictions over a sustained period of time,” said Swami Sivasubramanian, vice president of Amazon Machine Learning at AWS.