News for hackers. Stuff that matters.
Re:What Do You Believe But Cannot Prove? by speek on Saturday May 09, 2009 @02:41PM (#601235)
Computational AI will never get anywhere close to true Intelligence, but Neural Networks might get close.

Even better would be a hardware _and_ software approach, not a brute forcing approach.

Re:What Do You Believe But Cannot Prove? by saurabh on Saturday May 09, 2009 @03:24PM (#601290)
I have pondered over this for a while. When we say human intelligence or "true intelligence" as you say it, the fact is, it is full of irrationality and errors. I believe that computational AI will go farther than we can imagine for the simple reason that we cannot process information the way a computer can; its relentless.
Re:What Do You Believe But Cannot Prove? by aristus on Saturday May 09, 2009 @04:14PM (#601370)
Insisting that computer intelligence resemble human intelligence is probably the characteristic that the future will mock the most, like the way we mock primitive overdecorated Victorian machines.
Re:What Do You Believe But Cannot Prove? by 8plot on Saturday May 09, 2009 @05:37PM (#601471)
I too have been pondering this and I think the field is lacking the concept of artificial _imagination_. Maybe intelligence = knowledge + imagination, and since we already have massive artificial knowledge (wikipedia for example), AI is just missing the imagination to comprehend all this data.
Re:What Do You Believe But Cannot Prove? by liuliu on Saturday May 09, 2009 @03:00PM (#601262)
I believe complex combination of algorithms can be as intelligent as people other than neural networks.
Re:What Do You Believe But Cannot Prove? by endtime on Saturday May 09, 2009 @04:34PM (#601396)
What do you mean by "computational AI" and why doesn't it include neural networks?
Re:What Do You Believe But Cannot Prove? by speek on Monday May 11, 2009 @02:45PM (#604002)
Computational AI in that one tries to compute a top-down model of the mind.
Re:What Do You Believe But Cannot Prove? by endtime on Monday May 11, 2009 @05:32PM (#604368)
Hmm, okay. I don't know why you wouldn't consider bottom up models to be computational (and it doesn't seem to be a standard term; Google's second result is this comment thread), but at least I see what you meant now.