My First Experience with Coursera

Posted on September 13, 2012

2


CC-by-SA

It’s been a few weeks since I started with Dr. Ng‘s Machine Learning course at Coursera.org. It’s been a lot of work, but at the same time, it has been extremely rewarding.

For those who do not know what Coursera is, it is one of the several recently funded ventures that deliver college-level MOOCs (massive online open courses) on a range of subjects. My experience so far has been extremely positive.

The courses focus on practical applications of key, strategically selected courses. In the case of the Machine Learning course I’m taking, the typical laborious work of proving or deriving equations or algorithms is absent (though anyone with sufficient Calculus and CS background can reasonably discern where the formulas and algorithms are coming from.)

As far as I can see, it is possible to learn enough of a subject  in this venue (Machine Learning in my case) to do some productive “damage”. And that is my aim.

However, it requires discipline. The course provide a series of video lectures that are followed by sets of homework and/or programming assignments. It is easy to fall into the trap of simply watching the videos and doing little to no assignment.

But that is not learning. That is wasteful pseudo-edutaiment.

The key to get the most out of this course format (at least from my personal experience) is to do the homework first and foremost, and this is obvious. But most importantly, it is necessary to do the exams over and over.

Why is that? Well, one of the features of the Coursera medium is that enrolled students can take the tests several times up to a given deadline.  After the deadline, the highest score is taken as the official one for that test.

Coursera's Machine Learning, 2013-09-13

As you can see, you can take a test multiple times (up to a deadline)

Mucho importante. Every attempt provides a different set of questions. And here is the key, the importance of taking the tests over and over even if you have already scored the maximum.

By testing yourself against different sets of question, you reinforce what you have learned. What I do is that I print my test trials, and bind them next to the printed material for the course. I go over the failures, and I study them.

I guess every person is different, and so, each person has a different learning approach. However, it is my firm belief that this re-trial of tests is the key to learn the condensed material presented in these courses.

It is easy to test just to pass. And I’m not interested in just passing. If you want to really learn and internalize the  material, you need to be more proactive and aggressive with the course contents.

there is no try

Listen to Yoda

I for one cannot wait to take a few more courses (looking forward to Sedgewick’s Analytic Combinatorics I & II). Similarly, I cannot wait for a day when I can use whatever I’m learning here for a living 😉

CC-by-SA

Advertisements