Monday, March 3, 2014

Artificial Exigency



What is intelligence? Are Humans intelligent? If Humans are intelligent how do they create more problems for themselves than solve? If humans are not intelligent then how can they create something that is intelligent? What is Machine intelligence or artificial intelligence? Is Artificial intelligence about machines being made to do some so called intelligent operations or is it about machines developing ability to go beyond what they are programmed to do? Or is Artificial intelligence more about what humans want a machine to do?

So the question is that, can carrying out instructions as per a pre-determined program be called intelligence? Or can a machine perform actions beyond the program? If it does perform task beyond the program then, would we call it intelligence or an error? Infact even perfectly designed primitive machines sometimes do end up doing things that they were not programmed to do, would we say that the machine developed some intelligence of its own? 

I guess the core issue is to understand what intelligence actually is? It is only when we define what human intelligence is we could work towards developing a method of simulating it in machines. Intelligence has been a contentious issue as has been the various methods available to measure it. How many forms of intelligence exist? Are all or any of these forms of intelligence decided by genetics? Does intelligence has anything to do with memory and our ability to recall? Robert Sternberg and Howard Gardner came up with a list of types of intelligence. While Gardner had a longer list, I present the shorter list provided by Sternberg:

  • Analytical Intelligence: Intelligence that is assessed by intelligence tests.
  • Creative Intelligence: Intelligence that makes us adapt to novel situations, generating novel ideas
  • Practical Intelligence: Intelligence that is required for everyday tasks (e.g. street smarts)
Can any of this intelligence be created or developed or improved by exercising free will?  Are we stuck with what our genetics dictates? If we are stuck with what genetic dictates then we can't create intelligence we can only use the provided intelligence and we can only use to the extent provisioned.

So with such questions trained on human intelligence, there are similar queries that arise when we talk about artificial intelligence. Which form of intelligence are trying to create in a machine? I guess if it has to be a mix of all the three types. In the past few years we have seen machines that have demonstrated Analytical Intelligence, so I guess the real question is whether a machine could develop the other two forms. Would that mean that human can use his human intelligence to create a machine which is programmed for analytical intelligence but the the machine through some process develops a creative intelligence and practical intelligence to be really called intelligent? Or is it more a challenge for human intelligence to understand these two forms of intelligence and come up with code/method to design a machine that can demonstrate those two as well?


Often we hear statements such as ‘machines one day would overtake humans’ there would be a ‘fight between machines and humans in the near future’. How can a machine become more intelligent than what the humans created it to do? For a machine to develop an ability to think it has to be able to decide on its action irrespective of what it has been programmed to do, and the decided action still needs to be an acceptable, identifiable and valid action and not an aberration. So basically that means as explained above that, a machine needs to develop a free will and an ability to think to be truly intelligent beyond what a human has designed it to be. Can a machine develop consciousness too? What circumstances would lead a machine to develop this free will or this ability to think and decide on an appropriate response which it has not been pre-programmed to perform? 

Thus for me artificial intelligence, despite few decades of research and application, appears paradoxical. I also fail to understand the fascination towards hoping for a machine that one day would be more intelligent than humans and also behave and be like a human? However as a friend pointed out to me, probably it is man's way to try and ape nature and see if he could do better!

Labels:

0 Comments:

Post a Comment

Add your comments here

Subscribe to Post Comments [Atom]

<< Home

Random: Artificial Exigency

Artificial Exigency



What is intelligence? Are Humans intelligent? If Humans are intelligent how do they create more problems for themselves than solve? If humans are not intelligent then how can they create something that is intelligent? What is Machine intelligence or artificial intelligence? Is Artificial intelligence about machines being made to do some so called intelligent operations or is it about machines developing ability to go beyond what they are programmed to do? Or is Artificial intelligence more about what humans want a machine to do?

So the question is that, can carrying out instructions as per a pre-determined program be called intelligence? Or can a machine perform actions beyond the program? If it does perform task beyond the program then, would we call it intelligence or an error? Infact even perfectly designed primitive machines sometimes do end up doing things that they were not programmed to do, would we say that the machine developed some intelligence of its own? 

I guess the core issue is to understand what intelligence actually is? It is only when we define what human intelligence is we could work towards developing a method of simulating it in machines. Intelligence has been a contentious issue as has been the various methods available to measure it. How many forms of intelligence exist? Are all or any of these forms of intelligence decided by genetics? Does intelligence has anything to do with memory and our ability to recall? Robert Sternberg and Howard Gardner came up with a list of types of intelligence. While Gardner had a longer list, I present the shorter list provided by Sternberg:

Can any of this intelligence be created or developed or improved by exercising free will?  Are we stuck with what our genetics dictates? If we are stuck with what genetic dictates then we can't create intelligence we can only use the provided intelligence and we can only use to the extent provisioned.

So with such questions trained on human intelligence, there are similar queries that arise when we talk about artificial intelligence. Which form of intelligence are trying to create in a machine? I guess if it has to be a mix of all the three types. In the past few years we have seen machines that have demonstrated Analytical Intelligence, so I guess the real question is whether a machine could develop the other two forms. Would that mean that human can use his human intelligence to create a machine which is programmed for analytical intelligence but the the machine through some process develops a creative intelligence and practical intelligence to be really called intelligent? Or is it more a challenge for human intelligence to understand these two forms of intelligence and come up with code/method to design a machine that can demonstrate those two as well?


Often we hear statements such as ‘machines one day would overtake humans’ there would be a ‘fight between machines and humans in the near future’. How can a machine become more intelligent than what the humans created it to do? For a machine to develop an ability to think it has to be able to decide on its action irrespective of what it has been programmed to do, and the decided action still needs to be an acceptable, identifiable and valid action and not an aberration. So basically that means as explained above that, a machine needs to develop a free will and an ability to think to be truly intelligent beyond what a human has designed it to be. Can a machine develop consciousness too? What circumstances would lead a machine to develop this free will or this ability to think and decide on an appropriate response which it has not been pre-programmed to perform? 

Thus for me artificial intelligence, despite few decades of research and application, appears paradoxical. I also fail to understand the fascination towards hoping for a machine that one day would be more intelligent than humans and also behave and be like a human? However as a friend pointed out to me, probably it is man's way to try and ape nature and see if he could do better!

Labels: