So, Machine Learning, huh?
Yep. I've always been amused by the concept of teaching the Computer to do something. That's the sole reason I'm into programming (and because I believe programming is an art). And, Machine Learning takes it to a whole another level. In layman's term, Machine Learning is all about teaching a Computer to learn by itself. Whoa! that's super damn cool (You'll find only basic words in my vocabulary. I'm working on it though!). I mean think about it. You're teaching a thing (computer) to learn by itself. It sounds cool that way, but, underneath the hood it's all mathematics and statistics. I don't know how to explain myself about this. I really don't (I really need a work on my vocabulary). So, I'll just leave it to the masters.
But, why now?
- Data
- Raw Computing Power
- Algorithms
Without any of these, all the research on Machine learning was just theories. So, it's was just lurking in the background waiting for the right moment. Now, You'll all know that the computing power has grown exponentially in recent decades and we are today collecting tons of data for all kinds of weird things. All in all, we have all the basic needs to test all these theories and concepts that were developed, while it was in background. Today is the moment, Machine Learning researchers were waiting for. And they have garbed it tightly. Not only that, Researchers, today are trying to bring it more mainstream, to be able to accessible to everyone rather then being exclusive. Everyday, a new article pops out about the breakthroughs Machine Learning has achieved in various fields. I would say, Right Now is the best time to learn about Machine Learning.
Ok. So your Approach?
Now, this is were it gets a little tricky. Although, there are articles published on Machine Learning everyday, large number of them are Research papers which are difficult to grasp for an average student trying to get into this field. And, the introductory articles, although in abundance, are, meh!
And, there are no scarcity of introductory courses on machine learning claiming to teach you machine learning from scratch in about 3-6 months. Which is just plain impossible. You can learn to apply Machine learning principals in short amount of time, but, to learn or do research on Machine learning is a whole different thing. There's reason why it was exclusive to PhD.'s and Researchers.
My point is, if you want to apply Machine Learning principal's, you're golden. There's no scarcity of resources for that. You can be up and running in about a week to 3 months depending upon your background. Honestly, it's all a student need to/should learn if he/she wants to use machine learning in his/her respective field. The whole point is to make it more accessible so that anyone can apply the principals to his/her respective projects and do something which weren't possible in the recent past. But, If you want to go underneath the hood, learn the inner workings or want to implement the algorithms yourself & understand them, you'll have a hard time finding good quality material. You'll also need a solid background in statistics, linear algebra and little bit of calculus.
I've touched on a few courses and materials myself, and, was unsuccessful in learning from it (My inability, not the courses or materials' fault). And, to be honest, online learning which at first might sound easy and comfortable, is really, really difficult. It requires serious motivation to learn online alone (unless, you've someone with similar interest) sitting in front of computer. At least that's the case for me. And, after taking few courses I've realized that Top down approach is not for me (It may work for you perfectly). So, after lots of research and failed attempts to learn (I'm really determined on this) I've made a sort of course route for myself. Here's what it looks like:
And, this route will just teach me the basics to dig deeper into the field. We've just scratched the surface here. Machine learning is here to stay. In the next post I'll talk about the applications of machine learning and breakthroughs it has achieved so far. I'll also share all the resources I've gathered during my research on selecting a perfect top-down approach course route. I'm hopeful it shall help you. See you then. Peace!
- 0xelectron
P.S. So, I plan to update this blog regularly on my adventures (I can't make any promises, you know why, right?) in Machine Learning. I hope you'll join me.
The Game is On!
