CA Theory
- We have managed to get CAs to categorise a range
of items.
- However, we don't really understand much about them.
- It would be nice to leverage the attractor net theory,
but it's not clear how far that will go.
- It should (in no particular order):
- Describe when initial CAs form
- Be capable of largely stable long term memory
- but enable new memories to form: The Mule Problem
- Give an idea about net saturation
- Describe when CAs are forgotten
The Bicycle Riding Problem
- Show the breadth of a category (size of attractor basin)
- Say something about critical points in CA formation and
fractionation
- Say something about the size and behaviour of CAs
- Account for losing and gaining neurons and synapses
- Account for Hierarchical Categories
- Account for different types of CAs
