[bonds2034]

HELLO. ive been working on a theory for a new way to program artificial intelligence.
if we program in a constantly changing dictionary or words and actions. the dictionary will change as the input is givin. so say we program a computer to talk. and as it gets new information is will ask for a new definition or the word or theory and after some time of learning it will begin to write its own information in the dictionary as it receives new inputs by another program that simulates inquisitiveness. making it want to learn in a sense. and by giving it some form of wrong information for some things will provide it the sense of human intelligence. what are your ideas. i know my theory is somewhat flawed at this point but its really coming through. by having the constantly changing dictionary to me seams like the only way for it to be able to learn like a human. dont we store everything we know in somthing like a dictionary. please keep this thread going

[tylerfb11]

Your theory is simalar to cleverbot. only without the discrepincy of incorrect information. AI is a HIGHLY debated and contraversial topic within the programming community, ive been there before.

if you wish, you could email or pm me with deeper detail into your theory and i could look at it from a programmer's point of veiw.

[loserOlimbs]

There are a few chatbots that use this general approach actually.

[allnodcoms]

There is one fatal flaw in your plan... Reading, and being able to recite verbatim, the contents of a dictionary doesn't make you intelligent, it makes you "knowledgable".

Big difference.

Danny

[loserOlimbs]

Originally Posted by allnodcoms

There is one fatal flaw in your plan... Reading, and being able to recite verbatim, the contents of a dictionary doesn't make you intelligent, it makes you "knowledgable".

Big difference.

Danny

Thats more semantics than a flaw. If he can write it so that it passes a turing test, then it does meet the requirements of "Intelligent" as far as AI goes.

[allnodcoms]

Originally Posted by loserOlimbs

Thats more semantics than a flaw.

Not really. The Turing test is a test for intelligent responses, not knowledgeable ones. IF parsers and (as mentioned above) chat bots have passed Turing tests and no-one ever accused them of intelligence.

What the OP is proposing here is little more than a complex parser with a large database to draw pre-scripted replies from.

Danny