I recently read the book "Blink" by Malcom Gladwell. I highly recommend this book to everyone. It talks mainly about how our minds analyze situations quickly and how some of the conclusions we draw are highly accurate and some are horribly off the mark.
One such term he uses is "thin-slicing". This is understood to mean taking a tiny slice of an experience, sight, sound, or whatever and extracting a great deal if information out of that situation. An example is given about a statue purchased by some museum that was subsequently examined by a great deal of people and thought to be legitimate. Tons of analysis and scientific study was done and concluded the statue was exactly what was claimed. Then, the top expert in this field was brought in, and within 3 seconds of seeing the statue, he knew it was fake. He felt this. After more analysis, his intuition turned out to be true. People have these experiences daily, where they know something is true but can't explain why. This is what "Blink" is all about. The tagline is "The power of thinking without thinking".
At one point, Gladwell mentions that our thin-slicing ability is directly correlated to our experiences. Thus, our intuition is informed by our prior experiences within a certain area. This is the reason the expert was able to instantly realize the statue was a fake. What this means is that experiences train our intuition.
I should mention that this is science. This is not an attempt to justify people "feeling" that something is true (take, for example, the belief in God). What we're talking about are situations that can be repeatedly tested and experimented with such that truth is knowable. Science has shown us that our intuition is a biological mechanism (although not yet fully understood).
So, if our intuition can be trained, my thoughts drifted to artificial intelligence. I've been reading recently about artificial neural networks, which are basically nodes in a graph with different weights on each edge, all thrown together. They're used in very specific situations and can be very, very good at coming up with answers to difficult questions. However, they must be trained.
The similarity here is striking. If we can train these neural networks, have we come up with an explanation of how a huge portion of our brains work? Figuring out why an artificual neural network arrived at an answer is notoriously difficult to do. Is this the reason we cannot figure out how the human brain works (yet)?
Anyway, there's no point to this blog post. I just found the connection fascinating, the link between how our intuition appears to operate and how we can accurately train neural networks to answer very specific questions. Neural networks = homegrown intuition-boxes. Weird! Isn't science cool?
