Brain model may help build human-like robot
Can a computer mimic the way the brain works? |
Because of the emergent properties within the brain, it is very complex and hard to simulate artificially. But researchers are looking to simulate how the brain works so that human-like artificial intelligence can be created, along with a better understanding of damage to our own brains. At the University of Waterloo in Ontario, Canada, Chris Eliasmith led research on a brain model called SPAUN - the Semantic Pointer Architecture Unified Network.
SPAUN lives inside a computer and can view images with a camera-like eye and draw responses to questions. SPAUN works similar to the brain in that both a natural brain and SPAUN use a network of neurons that communicate by changing their voltages. Although SPAUN has only 2.5 million neurons within its programming (the human brain has about 100 billion), it can still mimic many functions of the brain - functions that until now were thought only humans could perform.
All of SPAUN's tasks involve numbers - including advanced pattern recognition. One major drawback of the machine is that it cannot operate in real time. For every one second, it takes 2.5 hours. Another flaw in the design is that SPAUN cannot learn any new tasks. Its knowledge is entirely hard-wired.
This article was particularly interesting to me because of its potential implications in the medical world. The research done to create SPAUN could also be useful to model brain damage. Eliasmith and his colleagues looked at what happens when neurons in a simulation get destroyed at the same rate as in humans as they age. If the model could help in the field of restorative therapy for patients with brain damage, that could be beneficial to many people who have faced brain damage from traumatic incidents. This is close to me because my father suffered brain injury during a hunting accident, during which his amygdala and hippocampus were damaged. If SPAUN could be used to find new ways to restore brain function, that would be revolutionary.
Link to the full article here
0 Comments:
Post a Comment
<< Home