Stanford - coursera : Machine learning


Machine learning is a branch of Artificial Intelligence and is the science of getting computers to act without being explicitly programmed. It is used on several domains, such as:

  • Image recognition
  • Optical character recognition
  • Sound recognition
  • Prediction
  • Spam filtering
  • Web search
  • And so many more...

This free online course is the first one I followed and a precursor of the coursera website. 


The course is 10 weeks long. Each week is composed of videos, quizz and octave exercices (automatically checked online). It took me about 5 hours per week to complete the course, composed of roughly 2 hours watching the videos and 3 hours answering the quizz and exercices.

Week 1

  1. Introduction
  2. Linear Regression with One Variable
  3. Linear Algebra Review

Week 2

  1. Linear Regression with Multiple Variables
  2. Octave Tutorial

Week 3

  1. Logistic Regression
  2. Regularization

Week 4

  1. Neural Networks: Representation

Week 5

  1. Neural Networks: Learning

Week 6

  1. Advice for Applying Machine Learning
  2. Machine Learning System Design

Week 7

  1. Support Vector Machines

Week 8

  1. Clustering
  2. Dimensionality Reduction

Week 9

  1. Anomaly Detection
  2. Recommender Systems

Week 10

  1. Large Scale Machine Learning
  2. Application Example: Photo OCR
  3. Conclusion


Though it is certainly quite a time investment, I cleary recommend following this course if, as I did, you are interested in the field but have no idea how it works. Videos are very clear, and the exercices are doable and a great complement to the course material. You learn to solve real world problems in this exciting field.

Moreover, I had a really great return on investment on this course, since I used this new knowledge to solve various contests, such as the USPTO challenge (optical character recognition, 6th place) and the Nasa Robonaut challenge (object recognition, 1st place).


comments powered by Disqus