ChangeOrange-CSDN博客

原创论文笔记：Secure Federated Matrix Factorization

1. Introduction In recommendation systems, two types of users’ private information are leaked in traditional MF（Matrix Factorization）： (i) users’ raw preference data (ii)users’ learned latent feature vectors Prior studies have studied privacy-preserving M

2021-03-03 16:25:34 816

原创论文笔记：Understanding Membership Inferences on Well-Generalized Learning Models

目录Introduction关于ML privacy risk，作者的思考方向与答案：GMIA--Generalized MIAContributionsAssumptiontheory Introduction 背景论文：Membership inference attacks against machine learning models 1.Membership inference attack (MIA) can succeed on overfitted models with only blac

2021-01-09 20:26:29 620 1

空空如也

TA创建的收藏夹 TA关注的收藏夹

TA关注的人

原创 论文笔记：Secure Federated Matrix Factorization

原创 论文笔记：Understanding Membership Inferences on Well-Generalized Learning Models

空空如也

空空如也

原创论文笔记：Secure Federated Matrix Factorization

原创论文笔记：Understanding Membership Inferences on Well-Generalized Learning Models