TY - JOUR
T1 - Scaling laws in learning of classification tasks
AU - Barkai, N.
AU - Seung, H. S.
AU - Sompolinsky, H.
PY - 1993
Y1 - 1993
N2 - The effect of the structure of the input distribution on the complexity of learning a pattern classification task is investigated. Using statistical mechanics, we study the performance of a winner-take-all machine at learning to classify points generated by a mixture of K Gaussian distributions (''clusters'') in RN with intercluster distance u (relative to the cluster width). In the separation limit u1, the number of examples required for learning scales as NKu-p, where the exponent p is 2 for zero-temperature Gibbs learning and 4 for the Hebb rule.
AB - The effect of the structure of the input distribution on the complexity of learning a pattern classification task is investigated. Using statistical mechanics, we study the performance of a winner-take-all machine at learning to classify points generated by a mixture of K Gaussian distributions (''clusters'') in RN with intercluster distance u (relative to the cluster width). In the separation limit u1, the number of examples required for learning scales as NKu-p, where the exponent p is 2 for zero-temperature Gibbs learning and 4 for the Hebb rule.
UR - http://www.scopus.com/inward/record.url?scp=0039677510&partnerID=8YFLogxK
U2 - 10.1103/PhysRevLett.70.3167
DO - 10.1103/PhysRevLett.70.3167
M3 - ???researchoutput.researchoutputtypes.contributiontojournal.article???
AN - SCOPUS:0039677510
SN - 0031-9007
VL - 70
SP - 3167
EP - 3170
JO - Physical Review Letters
JF - Physical Review Letters
IS - 20
ER -