Explore the critical aspects of drift detection in AI model deployment in this 34-minute conference talk from GOTO Copenhagen 2022. Learn why drift detection is essential for maintaining model performance, understand the statistics behind it, and discover practical implementation techniques. Gain insights from Thomas Viehmann, co-author of "Deep Learning with PyTorch," as he shares his experience with open-source tools like TorchDrift.org and DriftDash.de. Understand how to effectively monitor deployed AI models, ensure their continued reliability, and implement robust drift detection processes in real-world scenarios.
Overview
Syllabus
Intro
Agenda
Why detect drift?
Statistics of detecting drift
Practical drift detection
Conclusion
Outro
Taught by
GOTO Conferences
Related Courses
-
Monitoring Machine Learning Concepts
-
Build and operate machine learning solutions with Azure Machine Learning
-
Production Machine Learning Monitoring- Principles, Patterns and Techniques
-
Scouter Models for Concept Drift Detection in MLOps
-
An Introduction to Drift Detection
-
Protecting Your Machine Learning Against Drift - An Introduction
