{"title":"A Collaborative Training Algorithm for Multi-Sensor Adaptive Processing","authors":"Joel B. Predd, Sanjeev R. Kulkarni, H. V. Poor","doi":"10.1109/CAMSAP.2007.4498024","DOIUrl":null,"url":null,"abstract":"In this paper, we discuss a local message passing algorithm for collaboratively training networks of kernel-linear least-squares regression estimators. The algorithm is constructed to solve a relaxation of the classical centralized kernel- linear least-squares regression problem. A statistical analysis shows that the generalization error afforded agents by the collaborative training algorithm can be bounded in terms of the relationship between the network topology and the representational capacity of the relevant reproducing kernel Hilbert space; this is in contrast to related approaches which relate the similarity structure encoded in the kernel and the network topology. The algorithm is relevant to the problem of distributed learning in wireless sensor networks by virtue of its exploitation of local communication.","PeriodicalId":220687,"journal":{"name":"2007 2nd IEEE International Workshop on Computational Advances in Multi-Sensor Adaptive Processing","volume":"12 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2007-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"3","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2007 2nd IEEE International Workshop on Computational Advances in Multi-Sensor Adaptive Processing","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/CAMSAP.2007.4498024","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 3
Abstract
In this paper, we discuss a local message passing algorithm for collaboratively training networks of kernel-linear least-squares regression estimators. The algorithm is constructed to solve a relaxation of the classical centralized kernel- linear least-squares regression problem. A statistical analysis shows that the generalization error afforded agents by the collaborative training algorithm can be bounded in terms of the relationship between the network topology and the representational capacity of the relevant reproducing kernel Hilbert space; this is in contrast to related approaches which relate the similarity structure encoded in the kernel and the network topology. The algorithm is relevant to the problem of distributed learning in wireless sensor networks by virtue of its exploitation of local communication.