The multi-keyword sorted searchable encryption is a practical and secure data processing technique.However,most of the existing schemes require each data owner to calculate and store the Inverse Document Frequency(IDF...The multi-keyword sorted searchable encryption is a practical and secure data processing technique.However,most of the existing schemes require each data owner to calculate and store the Inverse Document Frequency(IDF)value,and then dynamically summarize them into a global IDF value.This not only hinders efficient sharing of massive data but also may cause privacy disclosure.Additionally,using a cloud server as storage and computing center can compromise file integrity and create a single point of failure.To address these challenges,our proposal leverages the complex interactive environment and massive data scenarios of the supply chain to introduce a fast and accurate multi-keyword search scheme based on blockchain technology.Specifically,encrypted files are first stored in an Interplanetary File System(IPFS),while secure indexes are stored in a blockchain to eliminate single points of failure.Moreover,we employ homomorphic encryption algorithms to design a blockchain-based index tree that enables dynamic adaptive calculation of IDF values,dynamic update of indexes,and multi-keyword sorting search capabilities.Notably,we have specifically designed a two-round sorting search mode called“Match Sort+Score Sort”for achieving fast and accurate searching performance.Furthermore,fair payment contracts have been implemented on the blockchain to incentivize data sharing activities.Through rigorous safety analysis and comprehensive performance evaluation tests,our scheme has been proven effective as well as practical.展开更多
现有多视角聚类算法存在:1)在学习低维表征的过程中无法准确捕获或忽略嵌入在多视角数据中的高阶信息和互补信息;2)未能准确捕获数据局部信息;3)信息捕获方法缺少对噪声点鲁棒性等问题.为解决上述问题,提出一种自适应张量奇异值收缩的...现有多视角聚类算法存在:1)在学习低维表征的过程中无法准确捕获或忽略嵌入在多视角数据中的高阶信息和互补信息;2)未能准确捕获数据局部信息;3)信息捕获方法缺少对噪声点鲁棒性等问题.为解决上述问题,提出一种自适应张量奇异值收缩的多视角聚类(multi-view clustering based on adaptive tensor singular value shrinkage,ATSVS)算法.ATSVS首先提出一种符合秩特性的张量对数行列式函数对表示张量施加低秩约束,在张量奇异值分解(tensor singular value decomposition,t-SVD)过程中能够根据奇异值自身大小进行自适应收缩,更加准确地进行张量秩估计,进而从全局角度精准捕获多视角数据的高阶信息和互补信息.然后采用一种结合稀疏表示和流形正则技术优势的l_(1,2)范数捕获数据的局部信息,并结合l_(2,1)范数对噪声施加稀疏约束,提升算法对噪声点的鲁棒性.与11个对比算法在9个数据集上的实验结果显示,ATSVS的聚类性能均优于其他对比算法.因此,ATSVS是一个能够有效处理多视角数据聚类任务的优秀算法.展开更多
文摘The multi-keyword sorted searchable encryption is a practical and secure data processing technique.However,most of the existing schemes require each data owner to calculate and store the Inverse Document Frequency(IDF)value,and then dynamically summarize them into a global IDF value.This not only hinders efficient sharing of massive data but also may cause privacy disclosure.Additionally,using a cloud server as storage and computing center can compromise file integrity and create a single point of failure.To address these challenges,our proposal leverages the complex interactive environment and massive data scenarios of the supply chain to introduce a fast and accurate multi-keyword search scheme based on blockchain technology.Specifically,encrypted files are first stored in an Interplanetary File System(IPFS),while secure indexes are stored in a blockchain to eliminate single points of failure.Moreover,we employ homomorphic encryption algorithms to design a blockchain-based index tree that enables dynamic adaptive calculation of IDF values,dynamic update of indexes,and multi-keyword sorting search capabilities.Notably,we have specifically designed a two-round sorting search mode called“Match Sort+Score Sort”for achieving fast and accurate searching performance.Furthermore,fair payment contracts have been implemented on the blockchain to incentivize data sharing activities.Through rigorous safety analysis and comprehensive performance evaluation tests,our scheme has been proven effective as well as practical.
文摘现有多视角聚类算法存在:1)在学习低维表征的过程中无法准确捕获或忽略嵌入在多视角数据中的高阶信息和互补信息;2)未能准确捕获数据局部信息;3)信息捕获方法缺少对噪声点鲁棒性等问题.为解决上述问题,提出一种自适应张量奇异值收缩的多视角聚类(multi-view clustering based on adaptive tensor singular value shrinkage,ATSVS)算法.ATSVS首先提出一种符合秩特性的张量对数行列式函数对表示张量施加低秩约束,在张量奇异值分解(tensor singular value decomposition,t-SVD)过程中能够根据奇异值自身大小进行自适应收缩,更加准确地进行张量秩估计,进而从全局角度精准捕获多视角数据的高阶信息和互补信息.然后采用一种结合稀疏表示和流形正则技术优势的l_(1,2)范数捕获数据的局部信息,并结合l_(2,1)范数对噪声施加稀疏约束,提升算法对噪声点的鲁棒性.与11个对比算法在9个数据集上的实验结果显示,ATSVS的聚类性能均优于其他对比算法.因此,ATSVS是一个能够有效处理多视角数据聚类任务的优秀算法.
