Et tu, Machine Learning?

The rationale behind the rapid rise of Machine Learning can be summarized in a simple analogy. Imagine you have just invented the Wheel. At first, the idea was to have a rolling mechanism to take your boulder from one place to another (lever!). Then, you realized you could put two of these things together and make a cart. Much farther downstream, came the use of potter's wheel, roller skates and a million uses in-between. Likewise, Machine Learning started off as an esoteric way for academicians to find meaningful patterns in scientific data. But soon, it turned into ML for housing, for language, for breaking down every problem into one that can be (if not now, then soon enough) solved by a computer - you just had to provide sufficient input data to 'train' the models. Like the now ubiquitous wheel before it, Machine Learning is all things to every thing. That's remarkable but also significantly broad that no one would complain about the lack of ambiguity.

I am working on my own series of articles that will take use cases engendered by ML and in doing so, enable a clear understanding of what problems ML can solve and what remain projects of the not too distant future.