 |
|
|
 |
|
|
|
|
Jshep89 |
| | | | | | posted 2/16/2012 00:33 |  | | | | | | | | | I'm still learning about AI, and was wondering if someone could explain this to me. Why is learning so important, if a computer can't take that knowledge and make a decision without human interaction? Also, how would you teach a computer capable of learning something like chemistry?
| | |
| | |
jeremy duncan |
| | | | | | posted 2/16/2012 02:21 | | | | | | | | | | There is two types of AI, one that can take input and output based on the input so the two meet at point A and seem logical.
chatbot gets input and is programmed that input goes to this point, then it acts to get to that point with the user.
Alan here at ai.com asks what music you like, then you answer and it says what music it likes.
Then there is AI that decides what the user is moving to and this is like the AI learning to do experimental chemistry or advanced conceptual math.
There is no defined point to meet so the AI has to decide at which point to meet the user.
if it gets it right it seems intelligent.
So your question is right is your talking abut a chatbot like Alan, but not like skynet in the terminator movies, skynet runs into Connor: two different types of AI.
| | |
| | |
walter_1927 |
| | | | | | posted 2/18/2012 22:10 |  | | | | | | | | | Yes, I think that learning is mandatory for a universally applicable A.I.system.
If you work only with algorithms, then the field of application is limited. The brain can respond only in one certain field. But if you make a learning system, then it can learn and be applied to any field. For details, see the link
Walter
| | |  http://www.intelligent-systems.com.ar/intsyst/proposedBrain.htm | | |
|
|
|
|
|
