There is a maniacal hype around Machine Learning these days, both in Silicon Valley and beyond. Larger companies are acquiring smaller startups claiming to do ML. SalesForce, for example, has had a dozen acquisitions in the last eighteen months. More than half of those acquired companies claimed to do ML Only one actually had a solid basis for making such a claim. This hype extends to software engineers, executives, entrepreneurs, venture capitalists, journalists and the general population. Google search for “machine learning” has tripled in the last 2 years. The number of New York Times articles discussing “AI” in 2016 quadrupled from 2009 when expressed as a percentage of all NYT articles.
I call this phenomenon Bladdapism. Let me explain.
First, did you notice how I casually switched from “machine learning” to “AI”? Are ML and AI synonyms? Let’s unscramble the letters of my newly-coined word: AI ML DS PA BD. Get my drift? It’s an alphabet soup! Let’s see if we can work through them.
AI = Artificial Intelligence
- a broad philosophical / mathematical concept of a machine that can reason / behave like a human
- the idea is old, but in its modern incarnation it dates back to at least Alan Turing’s work in the 1940s (played by Benedict Cumberbatch in The Imitation Game)
ML = Machine Learning
- a computer science discipline of (automatically) learning patterns from data
- like, give it a lot of examples of patient descriptions and a cancer yes/no diagnosis, and may be it can learn to diagnose a new patient
- this is a relatively new area of human knowledge and expertise with increasingly successful applications in domains as wide as speech recognition to sentiment analysis, machine translation of natural language to face recognition in images
DS = Data Science
- this extremely popular term was coined by a mathematician in the Obama White House
- I haven’t met any computer scientist who knows what it means
- for a lot of people it refers to the manipulation of (numeric) data that falls beyond the scope of Microsoft Excel
PA = Predictive Analytics
- again, not sure what it means, but it is apparently as good as sliced bread
- Deloitte tells me it is “examining raw data to draw conclusions about information” — go figure!
BD = Big Data
- when the data is too big to fit into one computer?
- in some usages, the term seems synonymous to ML
There you have the Wørd: Bladdapism. Have fun with it!