Continuing the topic of artificial intelligence based on brain-like network structure. Yes, we can potentially replicate the structure of the brain, and the steps in that direction are already made. This artificial memory-processing units will have the potential learn and recognise patterns and develop understanding of things like symbols, metaphors, art, indirect meanings etc.
However, the potential does not mean reality. How can a computer gain the same interpretations of symbols as human brain unless the machine experiences the same reality? Meaning the same senses, the same problems, the same needs as a human. I think it's an oversimplification to interpret human brain on its own, without the body, emotions, survival needs that brain is actually trying to solve all the time. And to take it further - do we actually need a machine that is an exact replication of a human? I guess not?
My thinking was that the whole point of human-like artificial intelligence is get us help in ambiguous environments, like in Big Data analysis, where it's actually not clear what the questions are, but the data is floating in large amounts that is too much for one human brain to absorb. Plus loads of information floating around seem to be useless - isn't it the greatest thing that the human brain can do (sometimes): to distinguish what is important and what not? And it seems that evolutionary it did so well, but now we maybe failing (I am referring to climate change etc). Maybe artificial brains can help us a bit by consolidating all the information and fishing out the patterns that we should look at in more details?
Another interesting thing of course, is the hierarchical neural networks - aren't we all now just neurons in the global brain created through Internet? And if we are - what are the implications of that? Can we benefit from that somehow?
I tend to like questions more than answers :) :) :)
My thinking was that the whole point of human-like artificial intelligence is get us help in ambiguous environments, like in Big Data analysis, where it's actually not clear what the questions are, but the data is floating in large amounts that is too much for one human brain to absorb. Plus loads of information floating around seem to be useless - isn't it the greatest thing that the human brain can do (sometimes): to distinguish what is important and what not? And it seems that evolutionary it did so well, but now we maybe failing (I am referring to climate change etc). Maybe artificial brains can help us a bit by consolidating all the information and fishing out the patterns that we should look at in more details?
Another interesting thing of course, is the hierarchical neural networks - aren't we all now just neurons in the global brain created through Internet? And if we are - what are the implications of that? Can we benefit from that somehow?
I tend to like questions more than answers :) :) :)
No comments:
Post a Comment