Most commonly used distributed machine learning systems are either synchronous or centralized asynchronous. Synchronous algorithms like AllReduce-SGD perform poorly in a heterogeneous environment, while asynchronous using parameter server suffer from 1) communication bottleneck at servers when workers many, and 2) significantly worse convergence the traffic to is congested. Can we design an algorithm that robust being efficient maintaining best-possible rate? In this paper, propose...

With the advent of cloud computing, it has become increasingly popular for data owners to outsource their public servers while allowing users retrieve this data. For privacy concerns, secure searches over encrypted motivated several research works under single owner model. However, most in practice do not just serve one owner; instead, they support multiple share benefits brought by computing. In paper, we propose schemes deal with preserving ranked multi-keyword search a multi-owner model...

With the advent of cloud computing, it becomes increasingly popular for data owners to outsource their public servers while allowing users retrieve these data. For privacy concerns, secure searches over encrypted motivated several researches under single owner model. However, most in practice do not just serve one owner, instead, they support multiple share benefits brought by servers. In this paper, we propose schemes deal with ranked multi-keyword search a multi-owner To enable perform...

With the development of cloud storage, more data owners are inclined to outsource their services. For privacy concerns, sensitive should be encrypted before outsourcing. There various searchable encryption schemes ensure availability. However, existing search pay little attention efficiency users' queries, especially for multi-owner scenario. In this paper, we proposed a tree-based ranked multi-keyword scheme multiple owners. Specifically, by considering large amount in cloud, utilize TF ×...

With the advent of cloud computing, more and people tend to outsource their data cloud. As a fundamental utilization, secure keyword search over encrypted has attracted interest many researchers recently. However, most existing researches are based on an ideal assumption that server is "curious but honest", where results not verified. In this paper, we consider challenging model, would probably behave dishonestly. Based explore problem result verification for ranked search. Different from...

Federated learning is a new type of distributed framework that allows multiple participants to share training results without revealing their data privacy. As privacy becomes more important, it difficult collect from owners make machine predictions due the lack security. Data forced be stored independently between companies, creating "data silos". With goal safeguarding and security, federated greatly expands amount data, effectively improving shortcomings traditional deep learning, bringing...

The e-healthcare cloud system has shown its potential to improve the quality of healthcare and individuals' life. Unfortunately, security privacy impede widespread deployment application. There are several research works focusing on preserving electronic record (EHR) data. However, these have two main limitations. First, they only support `black or white' access control policy. Second, suffer from inference attack. In this paper, for first time, we design an attack-resistant with...

High-speed Internet becomes more widely accessible, the cost of connecting devices is reducing, are being created with Wi-Fi abilities and sensors built into those devices, technology costs cost-effective. All these things creating a ``storm" for Things (IoT) era. With emergence Things, connect through internet, there many that providing data. The If want to share data aim achieving abstract knowledge, they need pass network. Some provide sensitive information can be time. We should protect...

Recently, Goutham et al. proposed an ID-based user authentication with key agreement on elliptic curve cryptography(ECC), which is suitable to be applied in client-serverenvironment. The scheme mainly negotiates a temporary session between two parties. However, we find that the contains some security flaws, such as incomplete anonymity, no provision for updating private and so on. In light of this, propose new version anonymous protocol used client-server environment, especially, calculation...

The cloud-assisted industrial Internet of Things (IIoT) architecture can sustain highly available computation and massive storage services for modern systems. When data owners store IIoT to remote cloud platforms, the security will face tough challenges. Cryptographic technologies endow an ability guarantee confidentiality. However, traditional encryption techniques make access control searching malfunctioning. Recently emerging attribute-based keyword search (ABKS) primitive achieves...

Rule mining, for discovering valuable relations between items in large databases, has been a popular and well researched method years. However, such old but important technique faces huge challenges difficulties the era of cloud computing although which affords both storage scalability: 1) data are outsourced to due explosion high management cost, 2) moreover, usually encrypted first before being privacy's sake. Existing privacy-preserving rule mining methods only assume distributed model...

Organizations and individuals nowadays are more willing to outsource their data save storage management costs, especially with the push of cloud computing which offers both computation scalability. However, data, once being released a server, is no longer under its owner's control, privacy security herein become primary concern. To this end, users usually encrypt private before outsourcing it, however makes mining be very thorny challenging as big encrypted. Under such new circumstance, we...

In the scenario of cloud computing, it is necessary to consider privacy protection audio content as well copyright and authentication simultaneously. this paper, we propose an encrypted multipurpose watermarking scheme based on storage computing. We firstly perform domain discrete wavelet transform obtain coefficients. By using spread spectrum method, embed robust fragile watermarks into different sub-bands. then algorithms extract from watermarked audio. Some experiments are conducted...

Searchable encryption enables users to perform a keyword-based search over encrypted data. Most of the existing searchable schemes are based on single cloud server model, which has some limitations, such as point failure and malicious insiders. In contrast, multicloud model higher data availability security. However, directly moving keyword would cause larger storage costs, security cannot be distinctly improved. addition, different from cloud, challenge is how effectively verify query...

A novel quantum group signature scheme is proposed based on Chinese Remainder Theorem (CRT), in order to improve the security of signature. The generation and verification can be successfully conducted only if all participants cooperate with each other message owner's arbitrator's help. parallel algorithm applied efficiently compare restored original message. All operations signing verifying phase executed circuits. It has a wide application E-payment system, Online contract, notarization etc.

Federated learning has become an effective method to realize collaborative among multiple data centers. However, the problem of privacy leakage not been perfectly solved. We proposed a scheme based on homomorphic encryption with random masks cope private caused by honest but curious participants. Through combining Paillier mask, security process is enhanced trivial computation and communication cost. Experimental results show that decryption time reduced more than half after using CRT...

Audio has become increasingly important in modern social communication and mobile Internet, e.g., the employment of voice messaging Apps. In scenario cloud computing, we need to consider privacy protection audio content integrity simultaneously. this paper, study application an encrypted fragile watermarking a privacy-preserving authentication system cloud. We firstly implement computation Mel-frequency cepstral coefficient domain (MFCC-ED) with additive homomorphic encryption. By performing...

With the continuous progress of society, economic level has also been greatly improved, and people are paying more attention to intelligent humanized management in their lives. At same time, environmental problems hidden safety hazards increasing, posing a serious threat human life property. People have higher pursuit for quality life. In order ensure that personal, property other related information not leaked, national security construction is affected, one important issues. However,...