RATES OF CONVERGENCE IN ACTIVE LEARNING

被引：78

作者：

Hanneke, Steve ^{[1
]}

机构：

[1] Carnegie Mellon Univ, Dept Stat, Pittsburgh, PA 15213 USA

来源：

ANNALS OF STATISTICS | 2011年 / 39卷 / 01期

基金：

美国国家科学基金会;

关键词：

Active learning; sequential design; selective sampling; statistical learning theory; oracle inequalities; model selection; classification; EMPIRICAL PROCESSES; SAMPLE MODULI; INEQUALITIES; BOUNDS;

D O I：

10.1214/10-AOS843

中图分类号：

O21 [概率论与数理统计]; C8 [统计学];

学科分类号：

020208 ; 070103 ; 0714 ;

摘要：

We study the rates of convergence in generalization error achievable by active learning under various types of label noise. Additionally, we study the general problem of model selection for active learning with a nested hierarchy of hypothesis classes and propose an algorithm whose error rate provably converges to the best achievable error among classifiers in the hierarchy at a rate adaptive to both the complexity of the optimal classifier and the noise conditions. In particular, we state sufficient conditions for these rates to be dramatically faster than those achievable by passive learning.

引用

页码：333 / 361

页数：29