As small children, we are able to understand language, to be able to tell the cap from a dog, and play small games like a game of catch. These are the three most important things humans find very easy to do. However, computers and robots cannot do this.
The computer scientists have been working for years on how to accomplish these feats in robotics and artificial intelligence
Computers can finally play games can win chess they can even be the world champion aware they cannot talk about chess or even learn backgammon. Needless to say, humans are very broad minded and flexible computers are not.
Computer sciences have finally figured out what they are doing wrong they have not been studying the human brain.
In 1950, Alan Turing conducted an experiment hiding a computer from plain view and as the computer talks to see if it would be able to be distinguished from the human voice.
It has been obvious through the years that is an extreme no. The human brain works in ways completely different from digital computers. In order to build an intelligent machine the human brain must be studied. However, can it be replicated?
A software platform has been built that simulates the human brain it uses Hierarchical Temporal Memory (HTM). The software be can be geared with software tools then be trained. The software learns in the same way children do.
Jeff Hawkins, inventor of the Palm Pilot, is the founder of Palm Computing, Handspring, and the Redwood Neuroscience Institute, has been working on this theory, and has developed software. If you would like to know the details, go to http://www.numenta.com/.
During this study of the human brain, it was discovered that there is one section in the brain that is capable of higher levels of thought and perception. This part of the brain called the neocortex is a large part of the brain and is responsible for hearing language music and motor control.
One of the most complex aspects of the brain is that it decides what goes to what. For example when you hear something, that goes to one part of the brain and when you see something that goes to another part of the brain. Hawkins is most confused at this aspect.
To Think as We Do
Hawkins has taken this knowledge and applied it to his software thus the software can learn on many levels, as does the human brain.
Hawkins has also taken the time to study what the brain learns, how it learns, and when to learn. When our eyes spot something unusual we pick it up and look at it in several different ways or we stop and stare studying the unknown object. During that time, our brain decides first what we are seeing and as the eye changes a sort of picture of the object is stored in our memory to recognize that the next time it is seen.
As you can see, the developers along with Hawkins have their work cut out for them. There will be a lot more studying of the human brain and a lot more trial and error with hopefully more trial than error. Nevertheless, there has been a lot of success while going down this path of study.
Sources Jeff Hawkins "Learn Like a Human" Spectrum URL: http://www.spectrum.ieee.org/print/4982
Saturday, July 25, 2009
Subscribe to:
Posts (Atom)
