mohr@cory.Berkeley.EDU (Gordon Mohr) (11/16/90)
-- Does anyone know if it's possible to simulate simple associative or back-prop neural nets within a spreadsheet? (Not that it'd be desirable or efficient.) It seems that at least some of the newest, most powerful spreadsheets for Macs & IBM-compatibles could handle NNs--at least the XOR problem. Just curious... Gordon Mohr mohr@cory.berkeley.edu
apindor@gpu.utcs.utoronto.ca (11/17/90)
In article <8953@pasteur.Berkeley.EDU> mohr@cory.Berkeley.EDU writes: >-- >Does anyone know if it's possible to simulate simple associative or >back-prop neural nets within a spreadsheet? (Not that it'd be >desirable or efficient.) > >It seems that at least some of the newest, most powerful spreadsheets >for Macs & IBM-compatibles could handle NNs--at least the XOR problem. > >Just curious... > >Gordon Mohr >mohr@cory.berkeley.edu Check 'Computers in Physics', May/June 1990, p.324 Andrzej Pindor University of Toronto Computing Services pindor@gpu.utcs.utoronto.ca
dfausett@zach.fit.edu ( Donald W. Fausett) (11/17/90)
Yes, it is doable. It has been done by one person whom I am aware of, Wade Shaw, at Florida Institute of Technology. He reports that it runs very slowly, but it allows him to import data from existing databases without reformatting, and it permits graphic displays of the results with no additional effort.