Read the sections on connectionism in [G$^+$98] and
be prepared to explain parts of it to the class:
Chapter 2, 34-49; Chapter 5, 135-136, 142-143; Chapter 6, 158, 160-163;
Chapter 10, 307-308.
Imagine that you are trying to convince someone that the best way to
create an artificially intelligent agent is to simulate the parallel
processing activity of the brain using broadly connectionist resources.
Think of a list of 5 major features that you would list as selling
points for a connectionist approach, and be prepared to expand upon
each item in your list.
Imagine that you are trying to convince someone that connectionist
approaches, though they may be useful for modelling the brain, have very
little role to play in the attempt to model and understand the mind. The
most important and interesting mental processes, you will argue, occur
slowly, in serial, and are best modelled using standard tools. How would
you try to make your case?
If you can, try to also read at least one of the following:
[Bec98,CS98,CC98]
W. Bechtel.
Connectionism and the philosophy of mind.
In William G. Lycan, editor, Mind and Cognition: an anthology.
Blackwell, Oxford, 1998.
BF 171 Min.
Paul M. Churchland and Patricia S. Churchland.
On The Contrary.
MIT Press, Cambridge, Mass., 1998.
BD 418.3 Chu. Also in SCIENTIFIC AMERICAN 1990 p. 26-31.
Paul M. Churchland and T. Sejnowski.
Neural representation and neural computation.
In William G. Lycan, editor, Mind and Cognition: an anthology.
Blackwell, Oxford, 1998.
BF 171 Min.