31.10.2024: Cross-Validation, Double Descent (Part 1)

This video is part of the Machine Learning series taught by Prof. Hamprecht at Heidelberg University during the winter term 2024/2025. 00:00:00 – Current (Research) Consensus on Double Descent 00:02:45 – Quick Recap of Last Lecture 00:03:15 – How to Pick a Good Hyperparameter – Cross Validation 00:15:40 – Bias-Variance Decomposition 00:16:45 – Example for Regression: Expected Mean Square Error 00:29:53 – Example for Regression: Visualization 00:33:11 – Double Descent: Classical vs. Modern Regime 00:43:10 – Correction to Mean Squared Error The lecture is part of the „Machine Learning and Physics“ course offered at Heidelberg University. Further materials and course details can be found at: https://sciai-lab.org/teaching/