For a long time now, I’ve been planning to start learning about the deep learning, but whenever I start reading about it, either through books such as The MIT Press Deep Learning book, or through some online video or MOOCs, I get bombarded with strange mathematical formulas and complex equations. Of course this complex mathematics is quite necessary to understand Deep Learning models but it also repels a beginner from learning about this field, as a beginner is more result oriented.
If I was a PhD student and I was researching to make a specific algorithm perform better in some way, sure I’d have to learn about all those equations and formulas, but the reality is, I’m not, for now. I am just a curious student, who’d like to get a general overview about how Deep Learning works and build some apps to test different models on the real world data.
But for a long time, most of the resources available to learn about Deep Learning were focusing on those PhD level students, quite boring and technical, with little to no practical projects whatsoever. So, whenever I tried starting this endeavor to learn about Deep Learning, just after a day or two, I’d get frustrated and abandon it altogether.
You can’t imagine my surprise when I learned about this course, fast.ai, taught by Jeremy Howard. The approach that they are using to teach Deep Learning is quite opposite to what’s being used elsewhere. Instead of diving deeper into the technicalities from the start, they are using a top-down code-first approach where they let the student experiment with different algorithms and techniques by implementing them to solve real world problems, like from Kaggle competitions etc, from the first lecture.
This is what I’ve been looking for all along. A way to ease into the Deep Learning field, experiment with different models, and when I get comfortable then delve deeper into the mathematics.
The main objective or the teaching strategy being followed by Jeremy in this course seems quite interesting and attractive but let’s see how effective it is and how well it is executed.
I’m hopeful that this is going to be an amazing experience for me, I’ll learn a lot and practice a lot. I’ll keep blogging about my experience taking this course, the things I liked, the things I didn’t, the problems I faced and solutions to those. Let’s see how it turns out.