Can I have a reference for this?? I have been studing AI for the last ten years and have yet to come across a machine that learns. Emultes behavior patterns (i.e. Expert Systems) maybe, but never learn.
Bear in mind that behavioural psychologists would often argue that all ''learning'' is merely ''emulating behavioural patterns''. Learning is such a vague term, it would be wise to use something more well-defined before deciding whether a computer is capable of it or not.
Graylien, I agree, just randomly wiping out 90% (_any_ 90%) of the population would not do anything for my hypothesis. If you look at what happens in this kind of catastophe you''ll find that x species are wiped out because they cannot survive in the new environment (lack of enough plant life, lack of light, lack of smaller prey) which opens up gaps in the hierarchy of life that creatures which survive can diversify into. In reality what changes is the shape of graph, a single local maxima stops being a local maxima because the predators for this species no longer exists and there''s more space to expand into niches. Catastrophes affect the environment which changes the ability to survive of all the species (most for worse, some for better) and that''s what wipes out various lines of evolution.
Physics can form bonds and break bonds, is that all there is to intelligence? Cool!
My point being that it is the interaction that makes the emergent behaviour we talk of as intelligent. Interaction of molecules, interaction of instructions and memory. It''s simple, it doesn''t need to be complex on the base level. It really, really doesn''t.
I wasn''t attacking anything anybody said. I was just pondering a fundamental question on whether or not those basic elements found in a computer, can possibly form what we feel as intelligent. I agree interaction is key, but we must be certain our foundation is firm.
By doing those four things with bits, well you''re reading this aren''t you?? So it''s not a matter of wether operating bits is ENOUGH, it''s HOW do we use those operations so that we can create artificial life . The big picture .
That''s the real problem - you can get the basic ideas in theory but beginning to implement such a thing is very complex. I think that the idea of a "matrix" / "bytecode" sort of intelligence is a good one, but (as somebody said) it needs to be thought about a bit more. Does anybody have any sort of ideas as to what else you could use? The system would have to include "senses" (input) "action" (output), and the reading / writing / development of the "memory" and "intelligence" of the creature. Some ideas:
· Real life modelling - create a complete massive world based on atoms and atomic properties, set it off processing, and wait for ages for life to form advantages: just like the real thing disadvantages: too much processing power, chances of life forming again very low, world would run slowly (to be accurate to any decent level)
· Self-Modifying Code (the next level) - create a basic intelligence system in a small program with the ability to modify "code" when something goes wrong, according to the problem (i.e. muscle and sensor on the tail, if the tail hurts, change the tail muscle code) advantages: might work disadvantages: self-modifying code hard to implement, not accurate, slow
· Neural Networks - giving a hand by supplying the real life way of emergent intelligence (intelligence as an emergent property of a connection of nerve cells) advantages: seems to be a lot of work in this field disadvantages: ???
Joviex you might want to pick up this book "Brainchildren" by Daniel C. Dennett, it has an example from the 60''s of how they tried to teach a computer to recognize the
interesting enough you said tried, not succeeded.
quote: problem, the turing machine and many others. Also if you can try and pick up "The Mind''s I" by the same author its another excellent book on the whole artifical intellegence field.
Already read it, still no examples of a working machine that has learned the difference between machine state and human state. Sorry, still need to prove that first statement to me