Off-Line Learning From Clustered Input Examples
msra(1996)
摘要
We analyze the generalization ability of a simple perceptron acting on a structuredinput distribution for the simple case of two clusters of input data and alinearly separable rule. The generalization ability computed for three learningscenarios: maximal stability, Gibbs, and optimal learning, is found to improvewith the separation between the clusters, and is bounded from below by the resultfor the unstructured case, recovered as the separation between clusters vanishes.The asymptotic...
更多查看译文
关键词
asymptotic behavior,generalization error
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络