Whenever you're ready, walk me through how you'd approach this.
What is regularization, what types are there, and why is it used?
Explain the purpose of regularization in supervised learning and compare L1, L2, and elastic net penalties. Discuss how regularization affects model coefficients, bias, variance, feature selection, and optimization. Demonstrate your explanation with a small scikit-learn implementation that selects a regularization strength using validation data.