- A new set of questions has been raised from this work.
- How can we build better CAs? For instance, ones that persist
like psychological memories.
- How can we implement better associative memory?
- How can we learn rules with variable binding?
- How can plans be learned?
- One that might work for an MSc thesis is: how can we
implement speech recognition?
- Another that might work for a thesis is: can this be implemented
in neural hardware (with Steve Furber).
- Of course any cognitive model (e.g. the Stroop task) would also
work as a thesis.