A short post about online educational resources on machine learning. Perhaps it is a sign of increasing popularity of the field that there are now several courses on machine learning accessible online and for free.
Out of Standford University comes Andrew Ng’s almost ‘legendary’ machine learning lecture presented on the Coursera platform. The class consists of video lectures, multiple choice style review quizzes and programming projects in Matlab. A discussion forum supplies student interaction that otherwise only a traditional university program offers. The class covers a good verity of supervised and unsupervised learning concepts, like neural nets, support vector machines, k-means and recommender systems.
Another online course is Berkeley’s CS188.1x Artificial Intelligence out of the EdX platform by Harvard and MIT. This course offers video lectures on topic like decision theory and reenforcement learning with programming assignments in Python.
A third offering in this direction is Sebasitan Thrun’s course on Artificial Intelligence presented at Udacity. Here the lectures are themed around applications to robotics on the example of Google’s driverless car . Udacity’s takes an interesting educational approach. It presents lectures in short sequences that are frequently interrupted by quick multiple choice style review questions. The purpose is to keep the students attention focused on the lecture and provide the feel of real teacher interaction. Similar to the EdX course programming projects are done in Python. Here the student can paste his code directly into the Udacity browser window for instant feedback, without actually having to run Python on the local machine.
Another platform to watch is Kahn Academy that originally started at the elementary to high school education but appears to be destined to expand into the university and professional level.
With the cost of learning at traditional universities exploding, at least in some parts of the world, it is fascinating to observe the formation of free or low cost online educational platforms. Possible revenue models for these platforms could be that lectures are going to be offered for free but a payment is requested for tests and certifications. There are already professional designations like the chartered financial analyst (CFA) or the FRM for risk management that are based on certifications outside the traditional university education setting. Interestingly a lot of development in the e-learning area comes out of the traditional schools, as a successful online program is certainly a great advertisement and recruitment tool by itself.