The specific mathematical tools that recur throughout large-scale training: tensors, SVD, probability, autodiff, optimisation, and numerical precision.
This chapter requires a subscription to access.
What you'll unlock:
1. Tensors, Matrix Multiply, and Linear Maps
2. Eigendecomposition, SVD, and Low-Rank Approximation
3. Probability, Entropy, and Cross-Entropy Loss
4. Automatic Differentiation and the Chain Rule
5. Gradient Descent, Adam, and Optimisation Landscapes