?url_ver=Z39.88-2004&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Adc&rft.title=A+connectionist+model+for+categorical+perception+and+symbol+grounding&rft.creator=Greco%2C+Alberto&rft.creator=Cangelosi%2C+Angelo&rft.creator=Harnad%2C+Stevan&rft.subject=Cognitive+Psychology&rft.subject=Neural+Nets&rft.description=Neural+network+models+of+categorical+perception+can+help+solve+the+symbol-grounding+problem+%5BHarnad%2C+1990%3B+1993%5D+by+connecting+analog+sensory+projections+to+symbolic+representations+through+learned+category-invariance+detectors+in+a+hybrid+symbolic%2Fnonsymbolic+system.+Our+nets+learn+to+categorize+and+name+50x50+pixel+images+of+circles%2C+ellipses%2C+squares+and+rectangles+projected+onto+the+receptive+field+of+a+7x7+retina.+The+nets+first+learn+to+do+prototype+matching+and+then+entry-level+naming+for+the+four+kinds+of+stimuli%2C+grounding+their+names+directly+in+the+input+patterns+via+hidden-unit+representations.+Next%2C+a+higher-order+categorization+(symmetric+vs.+asymmetric)+is+learned%2C+either+directly+from+the+input%2C+as+with+the+entry-+level+categories%2C+or+from+combinations+of+the+grounded+category+names+(symbols).+We+analyze+the+architectures+and+input+conditions+that+allow+grounding+to+be+%22transferred%22+from+directly+grounded+entry-level+category+names+to+higher-+order+category+names.+Implications+of+such+hybrid+models+for+the+evolution+and+learning+of+language+are+discussed.&rft.date=1998&rft.type=Conference+Paper&rft.type=NonPeerReviewed&rft.format=text%2Fhtml&rft.identifier=http%3A%2F%2Fcogprints.org%2F622%2F1%2Ficann9~1.htm&rft.identifier=++Greco%2C+Alberto+and+Cangelosi%2C+Angelo+and+Harnad%2C+Stevan++(1998)+A+connectionist+model+for+categorical+perception+and+symbol+grounding.++%5BConference+Paper%5D++++(Unpublished)++&rft.relation=http%3A%2F%2Fcogprints.org%2F622%2F