Posts Tagged ‘HPC’

IBM is going to try and pull off an age-old dream of neural computing. 
Utilising nanotechnology they will try to mimic the way that positive and negative reinforcement works on a neural network, but in a physical sense like a biological brain, not just a software model. This could potentially become very interesting, especially if it turns out to be a route to mimicking real brain behaviour. First of all that would finally verify that we have gotten something right in neurology (a field that seems to be lacking models above everything else) and second, it’s a lot easier to scale a computer then a pack of interconnected fat, hence Raymound Kurzweils dream of man-machine singularity could be one step closer.

In the more pragmatic end of the spectrum this could lead to real-time pattern recognition (or at least very high speed on a more complex system) something that we indeed could use very much at the LHC experiment at CERN, since selecting the “interesting” events out of 10^9 every second is kind of hard already.

Read the interesting article at BBC News