Scouter Models for Concept Drift Detection in MLOps
MLOps World: Machine Learning in Production via YouTube
Overview
Explore the critical topic of model drift monitoring in MLOps through this 19-minute conference talk from MLOps World: Machine Learning in Production. Dive into the concept of Scouter models, an innovative technique for identifying concept drift when true labels are unavailable in production data. Learn from Kumaran Ponnambalam, Principal Engineer - AI at Cisco Systems Inc., as he explains what Scouter models are, how to identify the right scouter feature, and best practices for building and managing these models. Gain valuable insights into addressing the severe limitation of drift monitoring when true labels are not accessible, making this talk essential for data scientists and MLOps engineers seeking to maintain the continued success of their models in production.
Syllabus
Scouter Models for Concept Drift Detection
Taught by
MLOps World: Machine Learning in Production