Machine Learning Yearning [New Book – Free Draft – Signup By Friday June 24th (2016)

Machine Learning Yearning by Andrew Ng.

About Andrew Ng:

Andrew Ng is Associate Professor of Computer Science at Stanford; Chief Scientist of Baidu; and Chairman and Co-founder of Coursera.

In 2011 he led the development of Stanford University’s main MOOC (Massive Open Online Courses) platform and also taught an online Machine Learning class to over 100,000 students, leading to the founding of Coursera. Ng’s goal is to give everyone in the world access to a great education, for free. Today, Coursera partners with some of the top universities in the world to offer high quality online courses, and is the largest MOOC platform in the world.

Ng also works on machine learning with an emphasis on deep learning. He founded and led the “Google Brain” project which developed massive-scale deep learning algorithms. This resulted in the famous “Google cat” result, in which a massive neural network with 1 billion parameters learned from unlabeled YouTube videos to detect cats. More recently, he continues to work on deep learning and its applications to computer vision and speech, including such applications as autonomous driving.

Haven’t you signed up yet?

OK, What You Will Learn:

The goal of this book is to teach you how to make the numerous decisions needed with organizing a machine learning project. You will learn:

  • How to establish your dev and test sets
  • Basic error analysis
  • How you can use Bias and Variance to decide what to do
  • Learning curves
  • Comparing learning algorithms to human-level performance
  • Debugging inference algorithms
  • When you should and should not use end-to-end deep learning
  • Error analysis by parts

Free drafts of a new book on machine learning projects, not just machine learning, by one of the leading world experts on machine learning.

Now are you signed up?

If you are interested in machine learning, following Andrew Ng on Twitter isn’t a bad place to start.

Be aware, however, that even machine learning experts can be mistaken. For example, Andrew tweeted, favorably, How to make a good teacher from the Economist.

Instilling these techniques is easier said than done. With teaching as with other complex skills, the route to mastery is not abstruse theory but intense, guided practice grounded in subject-matter knowledge and pedagogical methods. Trainees should spend more time in the classroom. The places where pupils do best, for example Finland, Singapore and Shanghai, put novice teachers through a demanding apprenticeship. In America high-performing charter schools teach trainees in the classroom and bring them on with coaching and feedback.

Teacher-training institutions need to be more rigorous—rather as a century ago medical schools raised the calibre of doctors by introducing systematic curriculums and providing clinical experience. It is essential that teacher-training colleges start to collect and publish data on how their graduates perform in the classroom. Courses that produce teachers who go on to do little or nothing to improve their pupils’ learning should not receive subsidies or see their graduates become teachers. They would then have to improve to survive.

The author conflates “demanding apprenticeship” with “teacher-training colleges start to collect and publish data on how their graduates perform in the classroom,” as though whatever data we collect has some meaningful relationship with teaching and/or the training of teachers.

A “demanding apprenticeship” no doubt weeds out people who are not well suited to be teachers, there is no evidence that it can make a teacher out of someone who isn’t suited for the task.

The collection of data is one of the ongoing fallacies about American education. Simply because you can collect data is no indication that it is useful and/or has any relationship to what you are attempting to measure.

Follow Andrew for his work on machine learning, not so much for his opinions on education.

Comments are closed.