在任意采样分布下使用加权迹范数进行学习

Jun, 2011

Learning with the Weighted Trace-norm under Arbitrary Sampling Distributions

Rina Foygel, Ruslan Salakhutdinov, Ohad Shamir, Nathan Srebro

TL;DR研究使用带权痕迹范数学习时的采样分布问题，提出强有力的学习保证方法，并证明其对实际测试的效果更好。建议使用经验分布加权，而非真实分布加权。

Abstract

We provide rigorous guarantees on learning with the weighted trace-norm under arbitrary sampling distributions. We show that the standard weighte