Stanford EE104: Introduction to Machine Learning

Lesson 1: Stanford EE104: Introduction to Machine Learning | 2020 | Lecture 1 - course information

This short lecture gives course information for EE104. It points to the syllabus and schedule at ee104.stanford.edu and explains how to follow along with the rest of the course.

Play lesson

Lesson 2: Stanford EE104: Introduction to Machine Learning | 2020 | Lecture - 2 overview

An overview of machine learning and the topics the course will cover. It introduces the main themes, including supervised and unsupervised learning, and sets up the material for later lectures.

Play lesson

Lesson 3: Stanford EE104: Introduction to Machine Learning | 2020 | Lecture 3 - predictors

This lecture introduces predictors, the functions that map inputs to outputs in a learning problem. It covers how predictors are defined and used as the basis for regression and classification tasks.

Play lesson

Lesson 4: Stanford EE104: Introduction to Machine Learning | 2020 | Lecture 4 - validation

Validation is the focus here. The lecture explains why we hold out data to estimate performance and how validation helps detect overfitting during model selection.

Play lesson

Lesson 5: Stanford EE104: Introduction to Machine Learning | 2020 | Lecture 5 - features

This lecture covers features and feature engineering. It discusses data standardization and how the choice of features affects what a model can learn.

Play lesson

Lesson 6: Stanford EE104: Introduction to Machine Learning | 2020 | Lecture 6 - empirical risk minimization

An introduction to empirical risk minimization. The lecture explains how learning is framed as minimizing a loss over training data and introduces parametrized predictors.

Play lesson

Lesson 7: Stanford EE104: Introduction to Machine Learning | 2020 | Lecture 7 - constant predictors

This lecture looks at constant predictors as a simple baseline. It shows how they behave under empirical risk minimization and why they are useful for comparison.

Play lesson

Lesson 8: Stanford EE104: Introduction to Machine Learning | 2020 | Lecture 8 - non quadratic losses

The lecture moves beyond quadratic loss to other loss functions. It discusses how the choice of loss changes the learning problem and the resulting predictor.

Play lesson

Lesson 9: Stanford EE104: Introduction to Machine Learning | 2020 | Lecture 9 - house prices example

A worked example on house prices. The lecture applies the ideas from previous lectures to a concrete regression problem and shows how the pieces fit together.

Play lesson

Lesson 10: Stanford EE104: Introduction to Machine Learning | 2020 | Lecture 10 - non quadratic regularizers

This lecture covers non-quadratic regularizers and their role in controlling model complexity. It explains how regularization can reduce overfitting and improve robustness.

Play lesson

Lesson 11: Stanford EE104: Introduction to Machine Learning | 2020 | Lecture 11 - neural networks

An introduction to neural networks within the course's framework. The lecture describes how neural networks fit into the predictor and empirical risk minimization view.

Play lesson

Lesson 12: Stanford EE104: Introduction to Machine Learning | 2020 | Lecture 12 - classifiers

This lecture introduces classifiers and categorical outputs. It explains how classification differs from regression and sets up the classification problems covered next.

Play lesson

Lesson 13: Stanford EE104: Introduction to Machine Learning | 2020 | Lecture 13 - erm for classifiers

Empirical risk minimization is applied to classifiers. The lecture shows how to define and minimize loss when the output is a category rather than a number.

Play lesson

Lesson 14: Stanford EE104: Introduction to Machine Learning | 2020 | Lecture 14 - Boolean classification

This lecture focuses on Boolean classification, where outputs take one of two values. It covers how these simple classification problems are formulated and solved.

Play lesson

Lesson 15: Stanford EE104: Introduction to Machine Learning | 2020 | Lecture 15 - multiclass classification

The lecture extends classification to more than two classes. It discusses how multiclass problems are set up and how predictors handle several categories.

Play lesson

Lesson 16: Stanford EE104: Intro to Machine Learning | 2020 | Lecture 16 - probabilistic classification

An introduction to probabilistic classification. The lecture describes how classifiers can output probabilities over classes instead of a single label.

Play lesson

Lesson 17: Stanford EE104: Introduction to Machine Learning | 2020 | Lecture 17-erm for probabilistic classif.

This lecture applies empirical risk minimization to probabilistic classification. It covers how loss functions are defined when the model outputs probabilities.

Play lesson

Lesson 18: Stanford EE104: Introduction to Machine Learning | 2020 | Lecture 18 - unsupervised learning

The lecture introduces unsupervised learning, where data has no labels. It describes the goals of unsupervised methods and how they differ from supervised learning.

Play lesson

Lesson 19: Stanford EE104: Introduction to Machine Learning | 2020 | Lecture 19 - principal components analysis

This lecture covers principal components analysis, a common unsupervised method. It explains how PCA finds directions of variation in data and how it relates to distances from a subspace.

Play lesson