When I started learning how to program a little over a year ago, I was driven by the motivation of gaining the ability to create. After gaining some proficiency and pushing code as a profession I have started to arrive at the conclusion that while programming in itself is an immensely useful skill it is also very time defective and most of the time wasteful. Working at a startup, I have always been plagued by the question of whether the work I am doing right know is actually going to be beneficial and useful to our users instead of vanity work that both adds no value to the product and wastes my time. Being a devotee of the principles of lean startup, I wanted to turn to the data and find a way to effectively let our analytics drive our development focus. But even then, it is hard to avoid many manual iteration cycles of trying out new ideas and experimenting.
Until recently, I had resigned to accepting fact that this process was just part of the whole process of tech entrepreneurship and that the answer was to simply become more efficient and only build what was necessary. That is until I read a book called, The Master Algorithm by Pedro Domingos. In this book, Domingos introduced me to the wonders of machine learning in fine yet accessible detail. Up until then I had not given much thought into how Google exploded onto the scene with accurate speech recognition, or how Facebook’s eerily accurate face recognition worked.
But what really got me was the idea the Machine Learning could even result in curing cancer and possibly all disease. Now that is some powerful stuff. So therefore I’ve decided to start deeper learning of the mysteries of Machine Learning. It seems like the problem most worth solving that could have the most positive impact for our race – okay and perhaps the most negative depending on who you talk to. One thing is for certain, this technology has the potential to revolutionize everything and I want to be apart of the revolution.
My strategy for learning quickly and effectively is simple. I will first learn the high level concepts of the field along with the history. Then I want to get into basic functional understanding and learn how to use a few libraries in order to accomplish practical tasks. I am debating whether it will be worth getting down into the low level details and understanding the math behind all the concepts. I want to learn enough to be able to effectively use the technology in creative entrepreneurial ways, but I don’t necessarily want to become a data engineer.
If you are also into Machine Learning please leave a comment on what inspired you to begin. Or more importantly, share some resources that you found useful in starting!