With the digitalization of various industries, combination cloud computing and industrial Internet Things (IIoT) has become an attractive data processing paradigm. However, cloud-assisted IIoT still challenging issues, including authenticity data, untrustworthiness third parties, system robustness efficiency. Recently, a lightweight certificateless signature (CLS) scheme for IIoT, that was claimed to address both been proposed by Karati et al. (2018). In this paper, we demonstrate CLS fails...

This paper proposes an end-to-end emotional speech synthesis (ESS) method which adopts global style tokens (GSTs) for semi-supervised training. model is built based on the GST-Tacotron framework. The are defined to present emotion categories. A cross entropy loss function between token weights and labels designed obtain interpretability of utilizing small portion training data with labels. Emotion recognition experiments confirm that this can achieve one-to-one correspondence categories...

Video Person Re-Identification (Re-ID) is a task of retrieving persons from multi-camera surveillance systems. Despite the progress made in leveraging spatio-temporal information videos, occlusion dense crowds still hinders further progress. To address this issue, we propose Temporal Correlation Vision Transformer (TCViT) for video person Re-ID. TCViT consists Attention (TCA) module and Learnable Aggregation (LTA) module. The TCA designed to reduce impact non-target by relative state, while...

Abstract In this paper, we propose a framework for lightning-fast privacy-preserving outsourced computation in the cloud, which refer to as LightCom. Using LightCom, user can securely achieve outsource data storage and fast, secure processing single cloud server different from existing multi-server model. Specifically, first present general LightCom under equipped with multiple Trusted Processing Units (TPUs), face side-channel attack. Under design two specified fast toolkits, allow...

Recently in the IEEE Transactions on Dependable and Secure Computing (doi: 10.1109/TDSC.2019.28976752019), Miao et al. proposed a novel construction of Privacy-Preserving Attribute-Based Keyword Search Shared Multi-owner Setting (ABKS-SM), which can delegate keyword search tasks to cloud server provider (CSP) without revealing any useful information. Although authors claimed that offline guessing attacks be resisted ABKS-SM scheme, we show this scheme indeed suffers from four types hence...

Deploying data storage and query service in an untrusted cloud server raises critical privacy security concerns. This paper focuses on the fundamental problem of processing conjunctive keyword queries over a privacy-preserving manner. Previous tree-based searchable symmetric encryption (SSE) schemes, such as <i>IBTree</i> <i>VBTree</i>, can process secure efficient way. However, these schemes cannot address &#x201C;Result Pattern (<i>RP</i>)&#x201D; leakage, which be used to recover keywords...

The unique urban form on the ground and “city overlap city” phenomenon occurring underground at Kaifeng city, Yellow River floodplain, is investigated. Archaeological data historical geographical analysis were used to study of surface remains. Primary collected from four 25 m long drill cores which obtained different locations city sedimentary cycles quantitatively divided-out dated. results show that evolution Kaifeng’s mainly occurred over periods, first was before 225 BC; second took...

In expressive speech synthesis, there are high requirements for emotion interpretation. However, it is time-consuming to acquire emotional audio corpus arbitrary speakers due their deduction ability. response this problem, paper proposes a cross-speaker transfer method that can realize the of emotions from source speaker target speaker. A set tokens firstly defined represent various categories emotions. They trained be highly correlated with corresponding controllable synthesis by...

Nowadays, data privacy is one of the most critical concerns in cloud computing, and many privacy-preserving distributed computing systems based on trusted execution environment (e.g., Intel SGX) have been proposed to protect user's during cloud-outsourced computation. However, these SGX-based solutions are vulnerable some traffic analyses, loading all tasks into enclave introduces much overhead for frequent EPC-paging. In this paper, we propose a T-SGX framework, which keeps confidentiality...

Curriculum learning, a machine training strategy that feeds instances to the model from easy hard, has been proven facilitate dialogue generation task. Meanwhile, knowledge distillation, transformation methodology among teachers and students networks can yield significant performance boost for student models. Hence, in this paper, we introduce combination of curriculum learning distillation efficient models, where help data aspects. To start with, aspect, cluster cases according their...

RNA-protein interactions (RPIs) play a very important role in wide range of post-transcriptional regulations, and identifying whether given pair can form or not is vital prerequisite for dissecting the regulatory mechanisms functional RNAs. Currently, expensive time-consuming biological assays only determine small portion all RPIs, which calls computational approaches to help biologists efficiently correctly find candidate RPIs. Here, we integrated successful computing algorithm, conjoint...

Secure outsourced middleboxes are deployed in network function virtualization services that detect malicious activities on communications, which provides privacy-preserving deep packet inspection (DPI) over encrypted traffic. To boost filtering efficiency of packets, the two-layer middlebox architecture has been adopted recent DPI systems. Nevertheless, state-of-the-art solutions based mainly suffer from two limitations: i) cannot support dynamic rule addition; ii) failed to inspect...

Container technology has been used for running multiple isolated operating system distros on a host or deploying large scale microservice-based applications. In most cases, containers share the same kernel with and other host, application in container can make calls of like normal process host. Seccomp is security mechanism Linux kernel, through which we prohibit certain from being executed by program. Docker began to support seccomp version 1.10 disables around 44 out 300+ default. However,...