A5080343833- IoT and Edge/Fog Computing
- Privacy-Preserving Technologies in Data
- Cloud Computing and Resource Management
- Stochastic Gradient Optimization Techniques
- Age of Information Optimization
- Mobile Crowdsensing and Crowdsourcing
- Infrastructure Maintenance and Monitoring
- IoT-based Smart Home Systems
- Advanced Computing and Algorithms
- Cryptography and Data Security
- Caching and Content Delivery
- Distributed systems and fault tolerance
- Advanced Graph Neural Networks
- Blockchain Technology Applications and Security
- Video Surveillance and Tracking Methods
- Distributed and Parallel Computing Systems
- Advanced Neural Network Applications
- Visual Attention and Saliency Detection
- Vehicle License Plate Recognition
- Brain Tumor Detection and Classification
Hong Kong University of Science and Technology
2024
University of Hong Kong
2024
ShanghaiTech University
2022-2023
Shandong University of Science and Technology
2023
Shanghai Institute of Microsystem and Information Technology
2022
University of Chinese Academy of Sciences
2022
Federated Learning (FL) has emerged as a fundamental learning paradigm to harness massive data scattered at geo-distributed edge devices in privacy-preserving way. Given the heterogeneous deployment of devices, however, their are usually Non-IID, introducing significant challenges FL including degraded training accuracy, intensive communication costs, and high computing complexity. Towards that, traditional approaches typically utilize adaptive mechanisms, which may suffer from scalability...
With the massive use of GPU, task scheduling under CPU-GPU clusters has become an indispensable research topic. Unlike existing models, we propose innovative framework that users offload their tasks in heterogeneous Edge Clusters (ECs) instead general-purpose CPU clusters. The takes full advantage GPU's powerful parallel computing capabilities. Specifically, decompose each user into sequential segments and segments, which can be offloaded to CPUs GPUs ECs, respectively. By dis-cretizing...
With the rapid development of edge data intelligence, task offloading (TO) and resource allocation (RA) optimization in multiaccess computing networks can significantly improve Quality Service (QoS). However, for online scenario, traditional methods (e.g., game theory numerical methods) cannot adapt to dynamic environments. Deep reinforcement learning (DRL) is applied adjust policy get long-term rewards. Nevertheless, since joint problem TO RA nonconvex NP-hard, existing DRL guarantee high...
Federated Learning (FL) has emerged as a fundamental learning paradigm to harness massive data scattered at geo-distributed edge devices in privacy-preserving way. Given the heterogeneous deployment of devices, however, their are usually Non-IID, introducing significant challenges FL including degraded training accuracy, intensive communication costs, and high computing complexity. Towards that, traditional approaches typically utilize adaptive mechanisms, which may suffer from scalability...
Partitioning and offloading the deep neural network (DNN) model over multi-tier computing units have been recently proposed to shorten inference time. However, state-of-the-art cannot adapt large-scale problems for streaming tasks because of its exponential complexity. Besides, as an essential kind DNNs, grouped con-volutional networks (GCNNs) has not explored yet. Motivated by above facts, in this paper, we concentrate on chained DNNs (CDNNs) GCNNs tasks. Consider a typical heterogeneous...
With the continuous development of network and communication technology, edge-cloud collaboration can combine advantages cloud computing edge to provide efficient safety monitoring services for factories. How realize collaborative security is an urgent problem. To this end, paper proposes a video redundant frame elimination mechanism images with higher possibility abnormal events reduce system computation overhead; designs difference threshold-based scheduling strategy automatically control...
On offshore rigs, the gangway between boat and docking row is critical to safety of workers' lives. In order avoid occurrence unrigged gangway, this paper proposes a SL-YOLOv5-based framework for detecting in berthing platoon scenario accurately quickly discover operation site without gangway. The method divided into three parts: firstly, boundary against extracted by edge detection algorithm obtain contour region; Then proposed SL-YOLOv5 identifies locates approaching obtains bounding box...
In the cloud data centers of large technology companies, there are various resources such as computing, communication and storage, which provide stable quality-guaranteed services for subscribers. To get a comprehensive understanding real load trends to observe characteristics user tasks consuming resources, we analyzed workload Microsoft Azure Virtual Machines (VMs). Specifically, all VMs in cluster within one month 2017 2019. The existing public provides complete information each schema....