> I had to re-read your site, there you wrote "every time a neuron fires   
   > at the same time as its target, a signal will be sent to nearby   
   > neurons suppressing their learning around that time. This is to   
   > prevent local neurons all learning to respond to the same things. The   
   > thresholds of the neurons in the system should be initialised so that   
   > in general only a small fraction of the nodes in the entire system   
   > should be firing at any one time perhaps <1% perhaps <<1%" - is that a   
   > thing you are referring to?   
      
   Yes. In particular the first half of that paragraph. i.e. the words up   
   to "same things".   
      
   > Meanwhile, do you have a lil bit more formal description, I am   
   > interested to code a little javascript applet to see how your thing   
   > works.   
      
   I don't have a ready made formal description - at the time I was   
   experimenting with different mechanisms for competition and   
   I have now forgotten which one worked best!.. I never did that many   
   experiments on it at the time - the computers were painfully slow.   
   And as you know, I never wrote up any results.   
   What is required is a mechanism to make sure that the nodes   
   don't all learn the same thing - this can be done in a variety of ways   
   and you will find you have some parameters to tweak in order to get   
   it working well.   
      
   I don't check this group so often so it may be better for you to   
   email   
   me directly (mick@reiss.demon.co.uk) to discuss fine details.   
      
   Mick.   
      
   [ comp.ai is moderated ... your article may take a while to appear. ]   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)