基于多示例学习的不经意传输协议安全性分析

不经意传输协议在执行过程中要能够同时保证发送方的隐私和接收方的隐私。本文提出一种基于多示例学习的模型,对不经意传输协议的安全性进行了分析。研究发现,在以下前提下利用此模型可以使不经意传输协议的发送方经过几次通信后算得接收方所选择的数据：（1）接收方的询问请求具有连续性;（2）发送方对接收方的选择兴趣有一定的先验知识。从而表明在不经意传输协议的执行过程中接收方的隐私会受到侵犯,不经意传输协议对接收方不是绝对安全的。

Analysis of Safety of Oblivious Transfer Protocol Based on Multiple Instance Learning

Oblivious transfer protocol should protect both the privacy of the sender and the privacy of the receiver.This paper suggests a model based on Multiple Instance Learning and analyzes the safety of oblivious transfer protocol.This model can help the sender figure out the choices of the receiver after some communications on the assumptions that： 1.The query of the receiver is consecutive;2.The sender has some previous knowledge about receiver＇s interests,which implies that the privacy of the receiver can be intruded and OT protocol is not absolutely safe for the receiver.