A5056420341- Software-Defined Networks and 5G
- Cloud Computing and Resource Management
- IoT and Edge/Fog Computing
- Caching and Content Delivery
- Interconnection Networks and Systems
- Privacy-Preserving Technologies in Data
- Network Traffic and Congestion Control
- Distributed and Parallel Computing Systems
- Advanced Memory and Neural Computing
- Advanced Neural Network Applications
- Age of Information Optimization
- Cryptography and Data Security
- Network Packet Processing and Optimization
- Mobile Crowdsensing and Crowdsourcing
- Advanced Optical Network Technologies
- IoT Networks and Protocols
- Network Security and Intrusion Detection
- Advanced Data Storage Technologies
- Statistical Methods and Inference
- Software System Performance and Reliability
- Energy Harvesting in Wireless Networks
- Cloud Data Security Solutions
- Peer-to-Peer Network Technologies
- Neural Networks and Applications
- Advanced Computing and Algorithms
Southeast University
2014-2025
China Three Gorges University
2023-2025
Yichang Central People's Hospital
2025
Southeast University
2022-2024
Chinese University of Hong Kong
2022
Zhejiang University
2021
Shanmuganathan Engineering College
2015
National Institute of Technology Tiruchirappalli
2010-2013
National Center for Nanoscience and Technology
2011
University of Hong Kong
2004-2006
Edge intelligence, as a prospective paradigm for accelerating DNN inference, is mostly implemented by model partitioning which inevitably incurs the large transmission overhead of DNN's intermediate data. A popular solution introduces multi-exit DNNs to reduce latency enabling early exits. However, existing work ignores correlation between exit settings and synergistic causing incoordination device-to-edge. To address this issue, paper first investigates bottlenecks executing in edge...
The Artificial Intelligence of Things (AIoT) is an emerging technology that enables numerous AIoT devices to participate in big data analytics and machine learning (ML) model training, providing various customized intelligent services for industry manufacturing. Federated Learning (FL) empowers applications with privacy-preserving distributed training without sharing raw data. However, due IoT devices’ limited computing memory resources, existing FL approaches cannot support efficient...
While many hardware accelerators have recently been proposed to address the inefficiency problem of fully homomorphic encryption (FHE) schemes, none them is able deliver optimal performance when facing real-world FHE workloads consisting a mixture shallow and deep computations, due primarily their homogeneous design principle. This paper presents FLASH-FHE, first accelerator with heterogeneous architecture for mixed workloads. At its heart, FLASH-FHE designs two types computation clusters,...
Elastic computing enables dynamic scaling to meet workload demands, and Remote Direct Memory Access (RDMA) enhances this by providing high-throughput, low-latency network communication. However, integrating RDMA into elastic remains a challenge, particularly in control plane operations for connection setup. This paper revisits the assumptions of prior work on high-performance computing, reveals that extreme microsecond-level optimizations are often unnecessary. By challenging conventional...
Advanced heart failure in patients with valvular atrial fibrillation (VAF) poses a significant threat to human health. Noninvasive assessment of left remodeling various pathological conditions is instrumental guiding clinical treatment decisions, evaluating efficacy, and predicting prognosis. The study enrolled 63 diagnosed mitral stenosis (MS), among whom 44 presented concomitant (AF) 19 had sinus rhythm. Left volume functional parameters were evaluated using real-time three-dimensional...
Cloud data centers, such as Amazon EC2, host myriad big applications using Virtual Machines (VMs). As these are communication-intensive, optimizing network transfer between VMs is critical to the performance of and utilization centers. Previous studies have addressed this issue by scheduling flows with coflow semantics or VM placement traffic considerations. However, been conducted orthogonally. In fact, two mechanisms mutually dependent, complementary degrees freedom independently turns out...
Federated edge learning (FEEL) is a promising collaborative paradigm, which employs devices (EDs) to train machine models for federation. It opens countless opportunities enable intelligence. The increasingly diversified demands intelligent services are driving the deployment of various federations at edge. Existing works on FEEL focus single federation and ignore inter-federation device competition intra-device resource allocation, hinders applications FEEL. To address this issue, article...
Summary With the rapid development of information technology, enormous volumes data are being generated by enterprises at all times. The management and storage these large‐scale have always been challenging enterprises. As usually shared among users in a collaborative manner, secure access performance 2 key concerns for However, current solutions fail to meet requirements since they suffer from following drawbacks: (1) do not support fine‐grained control cannot strict enterprises, (2)...
In this paper, we analyze the performance of pilot-assisted least square (LS) and minimum mean squared error (MMSE) channel estimators for orthogonal frequency division multiplexing (OFDM) systems with transmit antenna diversity. We first provide a design pilot sequences to simplify estimators. then (MSE) performance, study leakage effect. When tap is not sample-spaced, our analysis shows that power will leak only other taps same antenna, but also belonging antennas. The across antennas...
Fraud detection on multi-relation graphs aims to identify fraudsters in graphs. Graph Neural Network (GNN) models leverage graph structures pass messages from neighbors the target nodes, thereby enriching representations of those nodes. However, feature and structural inconsistency graph, owing fraudsters' camouflage behaviors, diminish suspiciousness fraud nodes which hinders effectiveness GNN-based models. In this work, we propose DiG-In-GNN, Discriminative Feature Guided GNN against...
The world is moving towards a new realm of computing such as Internet Things. Things, however, envisions connecting almost all objects within the to by recognizing them smart objects. In doing so, existing networks which include wired, wireless, and ad hoc should be utilized. Moreover, apart from other networks, network full security challenges. For instance, MANET (mobile network) susceptible various attacks in black hole its versions do serious damage entire infrastructure. severity this...
In recent years, deep neural networks (DNNs) have witnessed a booming of artificial intelligence Internet Things applications with stringent demands across high accuracy and low latency. A widely adopted solution is to process such computation-intensive DNNs inference tasks edge computing. Nevertheless, existing edge-based DNN processing methods still cannot achieve acceptable performance due the intensive transmission data unnecessary computation. To address above limitations, we take...
With the development of multi-queue data centers, various efforts have leveraged Explicit Congestion Notification (ECN) to achieve high throughput and low latency for communication. However, one deep-seated problems is that micro-burst traffic in networks could cause instantaneous queue length exceed ECN threshold, leading significant mismarkings ECN. Such further lead severe network performance degradation. In this paper, we propose Micro-Burst aware (MBECN+) mitigate issue. MBECN+...
It is challenging to allocate the network bandwidth virtual machines(VMs) hosting communication-intensive applications. Due temporal and spatial variability of hosted applications, it crucial how much be reserved for each VM when adjust it. Prior approaches typically resort predicting applications' demands, according which VMs are placed once all or periodically migrated. However, recent works conceded that demands applications can only accurately derived right before execution phase. In...
In Network Function Virtualization (NFV), multiple network functions cooperate to provide various services. To reduce the end-to-end latency through a chain of functions, research hotspots have turned complete NF parallelism frameworks. However, several issues remain in them such as manual dependency analysis on NFs and excessive for NFs. Therefore, this paper, we present ParaNF, an effective delay-balanced framework. ParaNF mainly consists two logical components. First, orchestrator...
Abstract Federated edge learning (FEEL) provides a promising device‐edge collaborative paradigm, which enables devices to parallel participate in model co‐creation while preserving user privacy, opening countless opportunities enable intelligence. With the growing demand for intelligent services, extensive FEEL deployment is inevitable. Nevertheless, existing FL schemes neglect two unique features (i.e., resource heterogeneity and data heterogeneity) real‐world thus may negatively affect...