Routing low-confidence loan scores to human review on AWS
AIF-C01
Guidelines for Responsible AI
Medium
Question 33 of 65
A lender uses an ML model to score loan applications. Company policy requires a person to review any score the model returns with low confidence before the application is decided, and the lender does not want to build its own reviewer queue and task interface.
Which AWS service will meet these requirements?
Show answer and explanation
BAmazon Augmented AI (Amazon A2I)
Amazon A2I provides human review of ML predictions and lets reviewers step in when a model cannot make a high-confidence prediction, and it removes the work of building the review system. Ground Truth also uses human workers, but it labels training data rather than reviewing live predictions. Model Monitor watches a deployed model for drift and routes nothing to a person. Mechanical Turk supplies a workforce but leaves the queue and task interface to build.
Source: docs.aws.amazon.com/
Anonymous, no account needed. Your tally stays in this browser.