Simple Rules with CAs
- The first thing that is new, is that we managed to
get the system to move onto new states.
- You can't do this with a Hopfield net, but
assymetric connections enable us to do it.
- We broke the system into four different nets.
- We set the inputs between networks according to
the image above, with arrows being excitatory
and forks inhibitory.
- The input to internal connections are just
a pass through. That is if we ignite
1 in input, 1 ignites in internal.
- The tricky thing is the rule. Here we have
to have the 1, 2, and + active before the
1+2=3 rule ignites.
- This rule then ignites the internal 3, and
turns off the 1, + and 2.
- Done, shuts the input and rule off, leaving internal
3 running.
- This is an example of changing state in a neural
system.
