Model Change Active Learning in Graph-Based Semi-supervised Learning

作     者：Kevin S.Miller Andrea L.Bertozzi Kevin S.Miller;Andrea L.Bertozzi

作者机构：Department of MathematicsUniversity of CaliforniaLos Angeles520 Portola PlazaLos AngelesCA 90095USA 

出 版 物：《Communications on Applied Mathematics and Computation》 (应用数学与计算数学学报（英文）)

年 卷 期：2024年第6卷第2期

页      面：1270-1298页

核心收录：

学科分类：12[管理学] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 081104[工学-模式识别与智能系统] 08[工学] 0835[工学-软件工程] 0811[工学-控制科学与工程] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：supported by the DOD National Defense Science and Engineering Graduate(NDSEG)Research Fellowship supported by the NGA under Contract No.HM04762110003 

主　　题：Active learning Graph-based methods Semi-supervised learning(SSL) Graph Laplacian 

摘      要：Active learning in semi-supervised classification involves introducing additional labels for unlabelled data to improve the accuracy of the underlying classifier.A challenge is to identify which points to label to best improve performance while limiting the number of new labels.Model Changeactive learning quantifies the resulting change incurred in the classifier by introducing the additional label(s).We pair this idea with graph-based semi-supervised learning(SSL)methods,that use the spectrum of the graph Laplacian matrix,which can be truncated to avoid prohibitively large computational and storage *** consider a family of convex loss functions for which the acquisition function can be efficiently approximated using the Laplace approximation of the posterior *** show a variety of multiclass examples that illustrate improved performance over prior state-of-art.