Neural Models
- The model that I and my group are exploring are fatiguing, leaky
integrators.
- There are lots of other people studying more refined models.
- There are people using less sophisticated models.
- We think these models are at the right level of sophistication
to be simulated on digital machines.
- We've managed to develop hierarchical categories, and associative
memories.
- We keep Hebb's concept of Cell Assemblies at the center
of our thinking.
- We have ideas about more complex structures (e.g. cognitive maps,
semantic nets), and processing via rules (variable binding)
and other mechanisms.
- We're also looking at implementation on parallel hardware to
start doing really complex processing.
- It's not a complete model, but we are looking to eventually
build a cognitive architecture based on this or a slightly
modified model.
