A5100454947- Distributed systems and fault tolerance
- Parallel Computing and Optimization Techniques
- Advanced Image and Video Retrieval Techniques
- Image Retrieval and Classification Techniques
- Advanced Data Storage Technologies
- Higher Education and Teaching Methods
- Robotics and Sensor-Based Localization
- Image Processing and 3D Reconstruction
- Software-Defined Networks and 5G
- Cognitive Functions and Memory
- Educational Technology and Assessment
- Remote Sensing and Land Use
- Video Surveillance and Tracking Methods
- Optimization and Search Problems
- Medical Image Segmentation Techniques
- Simulation and Modeling Applications
- Education and Work Dynamics
- Interconnection Networks and Systems
- Data Visualization and Analytics
- Online Learning and Analytics
- Education and Vocational Training
- Advanced Vision and Imaging
- Chaos-based Image/Signal Encryption
- Software System Performance and Reliability
- Advanced Steganography and Watermarking Techniques
Soochow University
2024
China University of Petroleum, East China
2007-2024
Affiliated Hospital of Qingdao University
2024
Qingdao University
2024
Civil Aviation University of China
2023
National University of Defense Technology
2023
Beijing Institute of Petrochemical Technology
2021-2023
Beijing Institute of Technology
2022
Guangzhou University
2022
Beijing University of Posts and Telecommunications
2018-2021
This paper presents nonblocking hash table algorithms that support resizing in both directions: shrinking and growing. The heart of the is a freezable set abstraction, which greatly simplifies task moving elements among buckets during resize. Furthermore, abstraction makes possible use highly optimized implementations individual buckets, including single flat array used for each bucket, improves cache locality.
The addition of transactional memory (TM) support to existing languages provides the opportunity create new soft- ware from scratch using transactions, and also simplify or extend legacy code by replacing synchronization with language-level transactions. In this paper, we describe our experiences transactionalizing memcached application through use GCC implementation Draft C++ TM Specification. We present recommendations that hope will guide effort integrate into languages, may contribute...
With the development of social life, prediction becomes more and important. As an emerging sequence modeling model, temporal convolutional network has been proven to outperform on tasks such as audio synthesis natural language processing. But it is rarely used for time series prediction. In this paper, we apply into problem. Gated linear units allow gradient propagate through unit without scaling so introduce in networks. order extract useful features, propose a multi-channel gated...
Time-based transactional memories typically rely on a shared memory counter to ensure consistency. Unfortunately, such can become bottleneck. In this article, we identify properties of hardware cycle counters that allow their use in place counter. We then devise algorithms exploit the x86 enable bottleneck-free runtime systems. also consider impact privatization safety and ordering constraints correctness, performance, generality our algorithms.
Community structure is an important characteristic of many real networks, which shows high concentrations edges within special groups vertices and low between these groups. related graph analysis, such as discovering relationships among communities, identifying attribute-structure relationships, selecting a large number with desired structural features attributes, are common tasks in knowledge discovery networks. The clutter the lack interactivity often hinder efforts to apply traditional...
Byte-addressable, non-volatile, random access memory (NVM) has the potential to dramatically accelerate performance of storage-intensive workloads. For applications with irregular data patterns, and that rely on ad-hoc structures, most promising model for interacting NVM is a transactional model. However, specifics matter significantly. We introduce two models programming persistent transactions. show how build concurrent from traditional software memories. then general model-specific...
Alloying is an effective way to modulate material's properties. In particular, graded alloying within a single domain of two-dimensional transition-metal chalcogenide (2D-TMC) great technological importance, for example, achieving band gap modulations. Here, we report facile method grow gradient Mo1–xWxS2 monolayers with large sizes and high crystal qualities via the chemical vapor deposition technique. The as-grown have composition W from x = ∼0 ∼1 in lateral dimension up 300 μm, span can...
Iterative, opportunistic and evolving visual sense-making has been an important research topic as it assists users in overcoming ever-increasing information overload. Exploratory visualization systems (EVSs) maximize the amount of can gain through learning have widely used scientific discovery decision-making contexts. Although many EVSs developed recently, there is a lack general guidance on how to evaluate such systems. Researchers face challenges understanding cognitive process supported...
Combining Network Functions Virtualization (NFV) with Software-Defined Networking (SDN) is an emerging and promising solution to provide scalable elastic network control service. In such a system, virtualized (NFs) need be consistently migrated from one instance another for various purposes, as resource optimization, fault tolerance, load balancing, etc. These migrations involve simultaneously coordinating updates the NF state SDN forwarding state. To solve this problem, we design two...
Recent microprocessors and compilers have added support for transactional memory (TM). While state-of-the-art TM systems allow the replacement of lock-based critical sections with scalable, optimistic transactions, there is not yet an acceptable mechanism supporting use condition variables in transactions. We introduce a new implementation variables, which uses transactions internally, can be used from within both sections, compatible existing C/C++ interfaces synchronization. By moving most...
This paper introduces a concurrent data structure called the mound. The mound is rooted tree of sorted lists that relies on randomization for balance. It supports O(log(log(N))) insert and O(log(N)) extract Min operations, making it suitable use as priority queue. We present two algorithms: first achieves lock freedom via pure-software double-compare-and-swap (DCAS), second uses fine grained locks. Mounds perform well in practice, support novel operations we expect to be useful parallel...
How to reduce the content placement cost of cloud delivery networks (CCDNs) is a hot topic in recent years. Traditional methods mainly by constructing trees, but they cannot adapt dynamic deployment proxy servers CCDNs. In addition, traditional method only provides paths according local decision-making without considering global dynamics congestion CCDNs, which also one main factors causing high placement. To solve these problems, we propose model based on Q-learning for called Q-content...
Objective This study aims to evaluate the effectiveness of Vaginal Microecology Evaluation System (VMES) in assessing dynamics vaginal microbiome (VM) throughout process vitro fertilization and embryo transfer (IVF-ET). Furthermore, it seeks explore potential correlation between distinct types VM ecology success rate IVF-ET. Methods employed VMES ascertain composition VM. Data were collected from infertile women who underwent their initial IVF-ET treatment for tubal factor January 2018...
Combining Network Functions Virtualization (NFV) with Software-Defined Networking (SDN) is an emerging solution to provide fine-grained control over scalable and elastic packet processing functions. Due changes in network policy, traffic characteristics, or physical topology NFV (SDNFV) systems, the controller needs carry out updates frequently, i.e., change data plane configuration from one state another. In order adapt a newly desired quickly, update process expected be completed shortest...
A selective encryption algorithm is proposed to improve the efficiency of high video coding (HEVC) and ensure security HEVC videos. The adopts integer dynamic coupling tent mapping optimization model as pseudo-random sequence generator, multi-core parallelization used generation mechanism. binstrings during process context adaptive binary arithmetic are selected for encryption, which conforms features invariable binstream compatible format in terms encryption. Performance tests six types...