Cognitive Modelling
- The parser is already a good cognitive model of a pretty
sophisticated task.
- We have unpublished work on the Stroop task, counting in
ICCM 2006, and associative memory submitted to ICCM 2009.
- We hope to do several other cognitive models.
- One advantage we have is biological time tied to neural behaviour.
- A particularly thorny issue is the behaviour of CAs.
- CAs are the bridge between neurons and psychology.
- As we, and others, have modelled them, they don't do a very good
job.
- They're pretty much on or off.
- A primary concern after we've finished CABot3 is an improved CA.
- One set if goals is to tie more and more behaviour to psychological
and neural data.
