TY - JOUR
T1 - Generalization performance of complex adaptive tasks
AU - Eisenstein, E.
AU - Kanter, I.
PY - 1993/6/7
Y1 - 1993/6/7
N2 - Optimal strategies for predicting correctly the output of a few new random inputs, when various feedforward networks are trained by noise-free random training examples, are examined analytically and numerically. The existence of a universal strategy for various generalization tasks is discussed, and indicates that the Bayes algorithm is not always the optimal strategy.
AB - Optimal strategies for predicting correctly the output of a few new random inputs, when various feedforward networks are trained by noise-free random training examples, are examined analytically and numerically. The existence of a universal strategy for various generalization tasks is discussed, and indicates that the Bayes algorithm is not always the optimal strategy.
UR - https://www.scopus.com/pages/publications/12044258413
U2 - 10.1103/physrevlett.70.3667
DO - 10.1103/physrevlett.70.3667
M3 - ???researchoutput.researchoutputtypes.contributiontojournal.article???
C2 - 10053932
AN - SCOPUS:12044258413
SN - 0031-9007
VL - 70
SP - 3667
EP - 3670
JO - Physical Review Letters
JF - Physical Review Letters
IS - 23
ER -