2005/11/12

Hypotheses about intelligence

Here's a set of hypotheses about intelligence:

Intelligence is similar to a muscle, it is a potential, a capacity.

Intelligence is the capacity of an individual to create and manage hypotheses. These hypotheses can be about anything, including itself.

Like a muscle, intelligence can be used and developed, or unused and undeveloped.

Unlike a muscle, intelligence is a component of the decision making process in an individual. In humans, decisions probably mostly involve emotions.

Intelligence can involve invalid hypotheses and thus can participate in the production of bad decisions. The key process that reviews fundamental hypotheses and creates new ones when needed is probably minimally involved in decisions. Day-to-day decisions are mostly taken using emotions and the existing set of hypotheses, and usually such decisions do not involve new fundamental hypotheses.

To this day, no mechanism has been created by man that could create a better set of hypotheses, in a non trivial situation.

When hypotheses are used as they are and not improved, then this process is probably similar to a mechanism (e.g., possibly one equivalent to predicate logic), albeit sophisticated.

There are two levels of managing hypotheses: (1) they are applied as they are, and (2) they are improved.

The process to improve hypotheses does attempt to keep consistency over the entire set.

This theory is applicable to other species than homo sapiens. It may be applicable to whales and dolphins for example.

Future intelligent robots will manage hypotheses, and once a machine will be able to improve sets of hypotheses reliably then machine intelligence will grow exponentially, far above human intelligence. Emotions in robots will probably be fake, i.e., simulated, to help communicate with humans.

Technorati:

Labels: , , , , , ,

0 Comments:

Post a Comment

<< Home