Thursday, September 07, 2006

Artificial intelligence ..

This is my chosen specialization for masters on and off you will be getting a lot of info regd this ..

Had my first ai class today and there was a very interesting notion which i would like to put in has always been thought to be the branch of computer science which mimics humans in their behaviour , the way they think , walk etc ..and acc to Professor Nikolaos this is major impediment in the functioning of tries to mimic human behaviour considering the fact that humans are superior in every aspect . There are certain circumstances wherein if you dont mimic the human way of doing things and use some other strategies you may find out that the device works better ...

Failures of ai

As i proceed further new and interesting relevations about ai are coming along my way is the more of the failures of ai rather than the success of ai that interest ..there was a time of excessive optimism in ai where in it was believed that ai would change the face of the world in ways similar to wat the computer or the internet have done but ai seriously fell short of these predictions ..many companies which were formed as a result of this optimism went bust ..the ones which managed to stay shifted to different lines now the question is whether all the research and work which goes into ai it all worth it ..are we not better of we the earlier systems we had ..well the ques has no clear cut answer rite now has too much of unpredictiblity to it ..we are not 100% sure of a systems response we cannot use ai to systems where human life is at stake ..aerospace or biomed etc ..
well as for the use of ai in gaming , entertaiment and intelligent interfaces is concerned ..that is a good field to venture into coz the risks involved are not as huge and ai does offer quite a lot of interesting features to these fields..

Japan - The Fifth generation revolution
Another factor that interests me is the fact that Japan has made significant strides in terms of implementation of ai as compared to the usa is as of now almost 5-6 years ahead of usa in tems of usa ..which incidently is the birth place of ai ...i read a very nice book by feigenbaum written in the 80's which warned researchers in usa about the fifth generation drive being undertaken by Japan and the consequences it would have ..but i guess at that people back in usa were to disappointed with ai to consider his warnings and the result is that rite now usa has lost out in the race to japan..

Control Systems vs AI
Remember KKM's aversion to all the ai methods being used . It seems this phenomenon is quite wide spread .Met my controls prof today and he was pretty surprised as to why am I taking the AI course . Acc to him ai is just a bunch of high expectations low result field . He says there is a point uptil which human intution can be used to solve a problem after that systematic methods lead to far greater results which we cannot see at the first instance . AI just basically relies on intuition and so cannot achieve any major results .

No comments: