.png){kind=icon}
返回
CONVEX ANALYSIS METHOD FOR DISTRIBUTED LEARNING ALGORITHM
.png){kind=icon}
PRE
AI
DOI：10.3934/mfc.2025035.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
. We provide a convex analysis method to bound the learning error of kernel regularized regression learning and distributed regression learning by combining the integral operator method, the spectral method, and the convex analysis method. We decompose the learning error as the sample error and the approximation error. The sample error is bounded by an improved convex analysis method whose advantage lies in separating the mean optimal solution from the empirical optimal solution, and the approximation error is presented with a K-functional. We provide an explicit learning rate which attains the classical optimal rate obtained by the integral operator method.
Keyword:
Distributed learning algorithm
reproducing kernel Hilbert space
con-vex analysis method
learning theory
期刊
M
IF:
0.8
论文数:
21
被引数:
0
机构
引用论文
.png){kind=icon}
Moduli of smoothness, K-functionals and Jackson-type inequalities associated with kernel function approximation in learning theory光滑模、K-泛函以及与学习理论中核函数逼近相关的Jackson型不等式
OAAI
OAAI分享收藏分享收藏
查看更多内容