Small World Topologies
- We have had some success with simple CA systems where
the CA is learned, and persists for realistic times.
- We have used small world topologies.
- There is some evidence that the brain is connected this way,
but it's not very solid.
- The idea is that some neurons are more central than others;
they are connected to more neurons, and are connected to
more central ones.
- The experiment presents neurons (like our categorisation work),
and the CAs form.
- Once formed they start to persist for reasonable times (seconds)
and then stop being active.
- Reactivation is in the right direction.
- We haven't fully explored reactivation.
- This is not published.
