Artificial Intelligence and Deep Learning For the Extremely Confused

Started by walkstall, April 01, 2018, 02:11:18 PM

Previous topic - Next topic

walkstall

A politician thinks of the next election. A statesman, of the next generation.- James Freeman Clarke

Always remember "Feelings Aren't Facts."

Solar

Fascinating. But it's not the AI that concerns me, it's the hybrid/AI that will be of issue in the future, and more than likely the demise of man or rather, the slavery of man.
Think about it, it's deep, but the repercussions are incalculable.
Official Trump Cult Member

#WWG1WGA

Q PATRIOT!!!

|Glitch|

There is a lot of misinformation about Artificial Intelligence (AI).  What we actually have today are knowledge-based expert systems, not AI.  Expert systems have existed since the 1970s.  They do one thing, and only one thing, very well.  For example, a number of universities have created medical diagnosis expert systems, that are used to diagnose medical issues and they all have a 90%+ success rate.  Those expert systems can't be used for any other purpose, other than what they were programmed to do.  That is not AI.

The difference between true AI and a knowledge-based expert system is whether or not it can do more than one thing.  Whether it is capable of going beyond its initial programming.  That is the true test of AI.  A medical diagnosis expert system is great for determining medical problems, but it can't be used to clean up an oil spill.  It is only good at doing what it was programmed to do, and nothing beyond that.  Whereas true AI is able to solve more than one problem, to go beyond its initial coding.

Being able to play chess really well is not AI, learning how to drive a vehicle is not AI, neural networks are not AI, and Deep Learning is not AI.