Machine Learning Foundations

Key Takeaways

• Learning rate selection: too large oscillates, too small converges slowly
• Stochastic vs mini-batch gradient descent trade-offs
• Convex vs non-convex optimization landscapes