Tuesday, June 28, 2005

Human intelligence, AI, and algorithms

I am interested in a word - one that can have scientific context, but certainly has a behavioral meaning.

Let's take the word "iteration" and link it to some interesting points - with some help from "Who's Afraid of Schrodinger's Cat," by Ian Marshall and Danah Zohar. Marshall is a British psychiatrist with a background in math and physics, and Zohar is an American physicist and philosopher (what I want to be in my next life).

In simple terms, an iteration is the repetition of a sequence of instructions or processes. When some experts on the brain look at cognitive processes - they think that the brain actually performs tasks based on this process, such as understanding language.

Okay, so stick with me...

Iterations can occur on the scientific side when a chemical reaction keeps producing a substance. In turn these actions create a stable structure - order out of chaos - we are talking the basis of life stuff. If this algorithmic process is repeated it generates shapes of "inexhaustible richness."

Marshall and Zohar say fractals (infinetly complex shapes that fall between the cracks of dimensionality and serve as the generating principles of chaos) that are then created imitate the patterns of nature, such as clouds, coastlines, interstellar dust and noise in electrical circuits.

Okay now to my point - it has been suggested that human intelligence is based on, or can be reduced to, a series of algorithms. And from here we get back to the artificial intelligence (AI) topic, because it may be algorithms that delivers AI capability.

So the question is can human intelligence be totally reduced to algorithms - or is it infintely more complex and able to "recognize shifting contexts and transcend paradoxes" making it difficult to mimic?

To be continued ...

No comments: