Machine Learning Notes

Module 01

Introduction to ML

What learning means and the shape of the field.

Module 02

k-Nearest Neighbors

The simplest classifier — and the vocabulary it forces us to invent.

Module 03

Linear Regression & Features

Fitting lines, and why the features often matter more than the line.

Module 04

Perceptron & Linear Separability

The first learning algorithm — and the things it can't do.

Module 05

Logistic Regression

Turning a line into a probability.

Module 06

Regularization

Why simpler models often generalize better.

Module 07

Support Vector Machines

The geometric idea of a margin, and the kernel trick.

Module 08

Decision Trees

Learning by asking questions.

Module 09

Hyperparameter Tuning & Validation

Choosing models honestly.

Module 10

Attribute Selection

Which features actually carry signal.

Module 11

Ensembles & Boosting

Why a society of imperfect learners outperforms a single one.

Module 12

Probabilistic Framework

Re-reading machine learning as inference.

Module 13

Dimensionality Reduction

Finding the few axes that matter.

Module 14

Unsupervised Clustering

Finding structure without labels.

Module 15

Introduction to Neural Networks

Connecting neurons together — and what changes when you do.

Module 16

Nonlinear Layers, Softmax & Cross-Entropy

The building blocks of every modern network.

Module 17

Autoencoders

Learning compact representations.

Module 18

Introduction to PyTorch

Building networks in code.

Module 19

Convolutional Neural Networks

Exploiting spatial structure in images.

Module 20

Recurrent Neural Networks

Learning over sequences.