Overview
Discover effective techniques for identifying silent failures in machine learning models in this 22-minute conference talk presented by Wojtek Kuberski from NannyML. Learn about the challenges of detecting subtle performance degradation in ML systems and explore practical strategies to monitor and maintain model reliability over time. Gain insights into advanced monitoring tools and methodologies that can help data scientists and ML engineers proactively address potential issues before they impact business outcomes.
Syllabus
How to Detect Silent Failures in ML Models - Wojtek Kuberski, NannyML [Presented in English]
Taught by
Linux Foundation