I’m back with another round of course notes, this time from Coursera’s Machine Learning, by Andrew Ng, as well as the entire Deep Learning Specialization sequence.

My history with these courses and material was curious. Much of the background mathematics, including calculus, I became extremely familiar with through the course of my physics degree years ago. Then, back around 2011, I went through the first run of Caltech’s Learning From Data online class which was, as is usual for Caltech, a tour de force of even more math, to really drill to the core of some of the most fundamental concepts in the field of machine learning. But at the time, it was obviously missing some of the more recent focus on neural networks, so I put in a mental bookmark to return to the subject and catch up on more recent developments once I had the opportunity.

That opportunity came recently, as a friend and I worked through the OG Machine Learning Introduction class on Coursera by Andrew Ng. This was a really great survey, but still felt a little dated, and certainly lacked many of the more recent developments in Deep Learning, so I still wanted more.

Done with the intro class, I started looking at Coursera’s Deep Learning Specialization sequence: a series of 5 courses, focused directly at deep learning techniques up to about 2017-2018.

The fateful moment came when I noticed that Coursera wanted a $49/month subscription¹ to access the course materials, but gave a 7-day “free trial” period.

Now, if you know me at all, you know that I don’t go for half measures of “slow and steady” progress. I’ve even written an essay on the subject called Slow and Steady is Bullshit.

So I saw the words “7-day free trial”, and read them as a challenge. All I’d have to do would be to exceed my usual pace of 1-week of a course per day, and accelerate it to 1-entire-course per day. Keep up that pace for five days, and I’d be done with time to spare.

If this sounds insane, that’s because it probably was (at least a little bit?). It was also only conceivable through the confluence of several factors: my previously-established deep familiarity with mathematics, my fluency with the python programming language (which was also the language of choice for the courses), the fact that the courses in the specialization sequence were shorter than is usual for semester-length classes (~2-4 weeks of material each, rather than ~12), and the browser console one-liner $('video').playbackRate = 3.5 to override the default maximum playback speeds of the lecture videos.

So I targeted a weekend, took a couple days off on either side, canceled or turned down all of my other plans, and got to work.

Fallback: this browser does not support PDFs. Please download the PDF to view it.

Download my course notes as a PDF.

Fallback: this browser does not support PDFs. Please download the PDF to view it.

Download my course notes as a PDF.

Long story short, I succeeded right on plan at a pace of 1 course per day, finishing the entire Deep Learning Specialization in 5 consecutive days. This was intense, and by the end I was literally dreaming about multiheaded attention and LSTMs. But, on the other hand, I’m now psyched and ready to dive into even newer developments in the field of deep learning since the 2017-2018 cutoff dates of the specialization.

I don’t know that I would recommend this extreme speedrun approach to others. In my case it was literally only possible because I already had a very solid background on all of the tools necessary for the course, and heck, it even helped that I had already spent a while getting used to the sound of Andrew Ng’s voice on significant speedup during the Machine Learning Intro class!

That said, if you want to learn through these courses yourself, figure out what sacrifices you can make, and then do it as fast as you can. Slow and steady is still bullshit.

  1. $49 / month is really a very reasonable amount of money to charge for access to the quality of material in the Deep Learning Specialization. No qualms there, and hats off to Coursera for providing such excellent lectures, notes, exercises, programming assignments, and exams.