Computers nowadays are capable of aiding us in many tasks. They even supersede and replace us in many of our daily jobs. But the Holy Grail in computer science hasn't yet been found; Artificial Intelligence (or AI from now on).
Yes I know computers can recognize and interpret our voice, yes I know they can recognize faces. But AI goes far beyond this. The "Intelligence" in AI suggests the computer is able to think for itself. This means it can observe a previously unknown situation, learn from it and adept to it. This sounds easy enough; Just program a computer to learn for itself. To understand what a difficult task this would be we'd have to understand the essence of learning.
Trying to grasp the essence of learning we must look at a human baby. They're born without control over their muscles, they cannot understand or speak words, they're unable to understand words and sentences. There's only instinct, a basic set of knowledge telling them, among other things, to suck at the breast. After they're born, babies have a whole lot of things to learn, some things they learn by seeing and trying. other things they discover themselves, yet again other things they learn by rationalisation and logical thinking. This brings us to a fundamental question. How do babies learn to learn?
So it turns out the essence of learning brings us to a paradox; learning to learn. If we want to teach machines to learn. We should teach them to learn learning.
How do we go about doing this? I'll leave that as an exercise for later;)
Posts
He expressed that the topic of whether machines could believe was excessively ambiguous, and proposed a test that was progressively explicit. artificial intelligence training in pune
ReplyDeleteThis is excellent information which is shared by you. This information is meaningful and magnificent for us to increase our knowledge about it. Keep sharing this kind of information. Thank you. Application Development Training Company Online
ReplyDelete