raja@msudoc.ee.mich-state.edu (Narayan S Raja) (05/19/88)
I'm posting this because all my attempts to e-mail steeg@ai.toronto.edu (and similar tries) bounced. Thanks for the info that a recent issue of AI magazine says something about Connectionism in expert systems. The article I referred to in my earlier posting (mentioning use of neural nets in mol. biology) was 'More human than ever, Computers are learning to learn', from NY Times, Sept. 15 1987. It said that at Los Alamos, a neural net was used to predict whether a given RNA sequence could code for protein synthesis, and that it 'performed better than conventional statistical techniques'. No references. Hope this helps, -- Raja.