Minor Stuff
- Symbol Grounding: I'm not saying that symbol grounding is a minor
problem, it's just that we've only made minor progress on it.
- My PhD student, Fawad Jamshed, is working on this.
- He's managed to do labelling and to learn five types of visual
objects. That gives you grounded symbols.
- He's now hoping to combine these grounded symbols and evaluate
sensibly.
- We've got a simulation of the Stroop task that is relatively accurate.
It just has the nets hooked up correctly.
- We'd like to modify that so it emerges from the
interaction of words and colours.
- We've also tested a modified fatigue model to get CAs that decay.
We use a combination of short and long term fatigue.
- However, it doesn't work with reactivated memories the way we'd like.
- Also, a modified decay model (decay decays) is similarly functional.
- This is work with Peter Passmore and seems like a nice way into the
CA3 problem.
