David Weinberger discusses deep learning, artificial intelligence and machine learning in "Everyday Chaos". His book "Too Big To Know" showed me some aspects of the internet and today's communication that helped me feel comfortable with what I know. The book on everyday chaos and complexity is doing the same thing.
These networks can be insanely complicated. For example, Deep Patient looked at five hundred factors for each of the hundreds of thousands of patients
Weinberger, David. Everyday Chaos (p. 54). Harvard Business Review Press. Kindle Edition.
In class on multiple regression, a statistical technique that investigates variables effects on another, we learned about all sorts of complexity, including how a pair of variables might affect a third. To really study the interactions of 500 variables, one could work with all the variables, all possible pairs of them, all possible triplets, etc. The total number of all the choices is 2 to the 500th power, a number with 150 zeros. When the author mentions "insane complexity", he isn't kidding.
With today's computers, researchers are accustomed to using 5 or maybe 10 predictor variables to understand the actions of a variable they are studying. Such a study uses multiple regression but there is a technique called "canonical correlation" where a set of predictors are used to predict more than one variable under study. Canonical correlation was too complex for most of our research to benefit but as computers, artificial intelligence and sets of information on hundreds of thousands of examples develop, we may be able to understand our world and ourselves Abetter than ever before.
Sometimes, the game called Go is considered among the oldest and also the most complex game in the world. Google engineers created a machine to play Go called AlphaGo. After a while, they build a 2nd version. They had the two machines play each other and version 2 beat the first one 100 times. Gigantic complexity, beyond what humans usually even contemplate, may be coming more under human understanding and control.
Our own brains can be considered insanely complex and humans may never come even close to understanding how they operate.
