?url_ver=Z39.88-2004&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Adc&rft.title=Pre-integration+lateral+inhibition+enhances+unsupervised+learning&rft.creator=Spratling%2C+M.+W.&rft.creator=Johnson%2C+M.+H.&rft.subject=Neural+Modelling&rft.subject=Computational+Neuroscience&rft.subject=Neural+Nets&rft.description=A+large+and+influential+class+of+neural+network+architectures+use%0Apost-integration+lateral+inhibition+as+a+mechanism+for+competition.+We+argue%0Athat+these+algorithms+are+computationally+deficient+in+that+they+fail+to%0Agenerate%2C+or+learn%2C+appropriate+perceptual+representations+under+certain%0Acircumstances.+An+alternative+neural+network+architecture+is+presented+in+which%0Anodes+compete+for+the+right+to+receive+inputs+rather+than+for+the+right+to%0Agenerate+outputs.+This+form+of+competition%2C+implemented+through+pre-integration%0Alateral+inhibition%2C+does+provide+appropriate+coding+properties+and+can+be+used%0Ato+efficiently+learn+such+representations.+Furthermore%2C+this+architecture+is%0Aconsistent+with+both+neuro-anatomical+and+neuro-physiological+data.+We+thus%0Aargue+that+pre-integration+lateral+inhibition+has+computational+advantages+over%0Aconventional+neural+network+architectures+while+remaining+equally+biologically%0Aplausible.&rft.date=2002&rft.type=Journal+(Paginated)&rft.type=PeerReviewed&rft.format=application%2Fpdf&rft.identifier=http%3A%2F%2Fcogprints.org%2F2380%2F1%2Fneurocomp.pdf&rft.identifier=++Spratling%2C+M.+W.+and+Johnson%2C+M.+H.++(2002)+Pre-integration+lateral+inhibition+enhances+unsupervised+learning.++%5BJournal+(Paginated)%5D+++++&rft.relation=http%3A%2F%2Fcogprints.org%2F2380%2F