The Framework and Questions
- This research fits in with an overall framework for developing
(real) AI.
- Make a computational model of human,
- using simulated neurons,
- that is a cognitive model,
- performs in an environment,
- and does interesting tasks.
- Repeat this developing more accurate models that do more interesting
tasks.
- This framework throws up questions along the way.
- One question is how do you learn a CA whose activation decays (over
seconds)like a biological memory?
- A second is how to learn rules with variable binding?
- A third is how do you do speech recognition?
- A fourth is how do you use neurons to categorise like people (submitted
EPSRC proposal)?
- A fifth is how do you use neurons to guide a robot?