Machine learning theory made easy.

So, you want to master machine learning. Even though you have experience in the field, sometimes you still feel that something is missing. A look behind the curtain.

Have you ever felt the learning curve to be so sharp that it was too difficult even to start? The theory was so dry and seemingly irrelevant that you were unable to go beyond the basics?

If so, I am building something for you. I am working to create the best resource to study the mathematics of machine learning out there.

Join the early access and be a part of the journey!

Math explained, as simple as possible.

Every concept is explained step by step, from elementary to advanced. No fancy tricks and mathematical magic. Intuition and motivation first, technical explanations second.

MSE figure
MSE figure

Open up the black boxes.

Machine learning is full of mysterious black boxes. Looking inside them allows you to be a master of your field and never be in the dark when things go wrong.

Black boxes figure
Black boxes figure

Be a part of the process.

This book is being written in public. With early access, you’ll get each chapter as I finish, with a personal hotline to me. Is something not appropriately explained? Is a concept not motivated with applications? Let me know, and I’ll get right on it!

Mathematics of machine learning timeline
Mathematics of machine learning timeline

The roadmap

Mathematics of machine learning book roadmap

This is what is covered in detail

Linear algebra

  • Vector spaces

  • Structure of vector spaces: norms and inner products

  • Linear transformations and their matrices

  • Eigenvectors and eigenvalues

  • Solving linear equation systems

  • Special matrices and their decomposition

Calculus

  • Function limits and continuity

  • Differentiation

  • Minima, maxima, and the derivative

  • Basics of gradient descent

  • Integration

Multivariable calculus

  • Partial derivatives and gradients

  • Minima and maxima in multiple dimensions

  • Gradient descent in its full form

  • Constrained optimization

  • Integration in multiple dimensions

Probability theory

  • The mathematical concept of probability

  • Distributions and densities

  • Random variables

  • Conditional probability

  • Expected value

  • Information theory and entropy

  • Multidimensional distributions

Statistics

  • Fundamentals of parameter estimation

  • Maximum likelihood estimation

  • The Bayesian viewpoint of statistics

  • Bias and variance

  • Measuring predictive performance of statistical models

  • Multivariate methods

Machine learning

  • The taxonomy of machine learning tasks

  • Linear and logistic regression

  • Fundamentals of clustering

  • Principal Component Analysis

  • Most common loss functions and what’s behind them

  • Regularization of machine learning models

  • t-distributed stochastic neighbor embedding

Neural networks

  • Logistic regression, revisited

  • Activation functions

  • Computational graphs

  • Backpropagation

  • Loss functions, from a neural network perspective

  • Weight initialization

Advanced optimization

  • Stochastic gradient descent

  • Adaptive methods

  • Accelerated schemes

  • The Lookahead optimizer

  • Ranger

Convolutional networks

  • The convolutional layer, in-depth

  • Dropout and BatchNorm

  • Fundamental tasks of computer vision

  • Alexnet and Resnet

  • Autoencoders

  • Generative Adversarial Networks

Want to find out more?

Listen to Practical AI’s interview with Tivadar about the book!

Practical AI 152: The mathematics of machine learning – Listen on Changelog.com

Help me write the book you want.
Get early access now!