Introduction to Deep Learning · HIT

Week 7   Part II · Training Infrastructure

Regularization & Generalization

Overfitting; dropout, weight decay, early stopping; basic data augmentation.

Curated, free, canonical references for this week: a course or lecture, a book chapter, a video, and an authoritative blog post or official tutorial. Each opens in a new tab.

Course
Stanford CS231n: Neural Networks Part 2 (Regularization) cs231n.github.io

Dedicated sections on L2/L1 regularization and dropout, plus train/val/test discipline.

Book
Understanding Deep Learning (Simon Prince), Chapter 9: Regularization udlbook.github.io

Covers overfitting, weight decay, dropout, early stopping, and augmentation as one toolkit (free PDF).

Video
StatQuest: Regularization Part 1, Ridge (L2) Regression youtube.com

Intuitive introduction to L2/weight-decay regularization and the bias-variance trade-off.

Blog / Docs
Dive into Deep Learning, 5.6 Dropout d2l.ai

A runnable tutorial explaining dropout, with both a from-scratch and an nn.Dropout version.

← Back to the Week 7 lab

PreviousWeek 6: OptimizationNextWeek 8: Convolutional Networks I