Human-in-the-Loop

Human-in-the-loop describes the deliberate inclusion of people in automated decision or learning processes to improve quality, robustness and accountability. Common in ML workflows for labeling, review and error correction. It defines roles, feedback loops and interfaces between human actors and systems. Implementation requires escalation, training and performance measurement.