comprehension in artificial intelligence!!!

Well, i am an outsider actually i am thorough with c and c++ so you guys may think what the wonder i am doin here, am extremely sorry for my uninformed intrusion actually i belong to c forum but recently saw abt AI and came to know something is the followuing true::



One of the biggest difficulties with AI is that of "comprehension".
Many devices have been created that can do amazing things,
but critics of AI claim that no actual comprehension by the AI machine has taken place.

can anyone plz tell me what this comprehension means in AI and what the most advanced AI machines built ever have not been able to perform this comprehension task..........


seeking for a kind reply,

nishant tiwari....

There probably isn't a universal answer on this, but I think that would refer to the fact that regardless of what AI has been created so far, the computer has never been aware of what it's doing. The AI can do nothing other than what it's programmer specifies. Even in machine learning, programmers are the ones who specify what to learn, and how to learn it.

Take a machine like Deep Blue, who can basically beat the World Chess Champion a little more than half the time. Although it is an incredible computational marvel (as it is said it could foresee up to 42 moves in advance) it lacked the slightest bit of intuition, which is seen when it loses to the world champion (who thinks about 12 moves in advance, but along more specific paths). So in the end, it's hard to tell if it's intelligent or not.

and it couldn't make a cup of tea if its life depended on it...

"Comprehension - AI"
This is really a very broad question and as happy_reaper put it giving
a universal answer is not possible. It involves questioning of the very
two fundamental aspects of the question 'comprehension' and 'intelligence'.
To 'comprehend' something is to 'understand' something and what exactly is understanding?
What do we mean when we say we have 'understood' something.
Grass is green or 1+1=2 . We say we understand both the statements but what does it mean?
May be association with some familiar elements or an intuitive feel of 'understanding...
Now how exactly a machine comprehends ? or learn ? or we are attempting to make it comprehened or make it learn?
We try to mimick our behaviour..
Humans in general have various response to a stimulus.Most of which follow a standard pattern. Thus we can generate a set of functions and say that a machine is nothing but a set of functions. You give a input - stimulus and you get a output- response...Thus when the machine encounters a similar situation it 'understands' by correlating and thus generates the response..Though what i have said is gross oversimplification this may roughly give you an idea..how the machine is trained or made to'comprehend'....

thanx............thanx a lot to all of you!!!



nishant tiwari

Many AI algorithms can claim to "learn" from different sets of inputs, (based on "learning rules"), but can they learn new "learning rules"? Its this sort of "meta" knowledge, or "feedback" that many may consider comprehension or intelligence, not just knowing, but knowing why you know. This is a huge step up in complexity, but nonetheless not impossible (comparable to the intelligence of a rat vs a human). I'd say right now, the field of AI can maybe effectively emulate the intelligence of the former, and given time, will eventually be able to surpass the latter.