Introduction
Machine learning (ML) systems have become ubiquitous across industries, from healthcare diagnostics to industrial process control. As these systems move from research prototypes to production deployments, the need for rigorous measurement and evaluation grows. Evaluation serves two intertwined purposes: …