[comp.ai.neural-nets] Citation request

pwm@unix.cis.pitt.edu (Paul W. Munro) (05/15/91)

I am looking for a paper (or papers) that investigate 
different stategies for selecting patterns under backprop
and the corresponding effect on network performance, convergence 
time etc.  I seem to recall a paper (by Baum?) in which networks
request patterns that lie near the category boundary...

Thanks in advance.

Please email directly to munro@lis.pitt.edu

- Paul Munro