Article
Computer Science, Information Systems
M. Sunita, Sujata V. Mallapur
Summary: This study proposes a novel BGP anomaly detection model that improves accuracy and reliability by extracting multiple features, using DBN for anomaly detection, and employing a hybrid RHMFO optimization for enhanced classification accuracy. The results show a significant improvement in accuracy compared to other models.
MULTIMEDIA TOOLS AND APPLICATIONS
(2023)
Article
Computer Science, Artificial Intelligence
Yiping Yan
Summary: This paper studies the application of the binding model algorithm in BGP anomaly detection and proposes two methods for detecting anomalies. It also introduces Cache 5G multimedia technology and a new type of Cache 5G multimedia technology called NCLMBCN. The NCLM BCN solution is used in the relay transmission system to improve the reliability of user reception by eliminating interference caused by relay transmission using cache 5G multimedia with users and repeaters.
Article
Computer Science, Hardware & Architecture
Marcelo Bagnulo, Alberto Garcia-Martinez, Stefano Angieri, Andra Lutu, Jinze Yang
Summary: This paper presents ASIRIA, a mechanism for detecting and avoiding route leak events on the Internet. By utilizing AS relationship information inferred from the Internet Routing Registries, ASIRIA provides immediate benefits to early adopters and can detect a majority of leakage events solely using currently available information.
Article
Computer Science, Information Systems
Cong Dong, Jiahai Yang, Song Liu, Zhi Wang, Yuling Liu, Zhigang Lu
Summary: Lateral movement plays a crucial role in network attacks, and it is important to identify traces of lateral movement to protect enterprise resources. Previous studies have proposed methods, but they often fail to raise high-quality alerts, making it difficult for security operators to respond to real threats in a timely manner. In this paper, we propose a novel approach called Behavior Deviation Measurement (BEDIM) to raise effective alerts for lateral movement detection. BEDIM can locate unusual connections and filter false alerts, and experiments show its effectiveness and accuracy.
COMPUTERS & SECURITY
(2023)
Article
Computer Science, Artificial Intelligence
Yiyue Li, Qicheng Lao, Qingbo Kang, Zekun Jiang, Shiyi Du, Shaoting Zhang, Kang Li
Summary: This paper proposes a universal unsupervised anomaly detection framework SSL-AnoVAE, which utilizes a self-supervised learning module to provide more fine-grained semantics for detecting anomalies in retinal images. Experimental results demonstrate the effectiveness of our proposed method for unsupervised anomaly detection, staging, and segmentation.
MEDICAL IMAGE ANALYSIS
(2023)
Article
Geochemistry & Geophysics
Chein- Chang, Chien-Yu Lin, Pau-Choo Chung, Peter Fuming Hu
Summary: This article introduces a new approach called iterative spectral-spatial hyperspectral anomaly detection (ISSHAD), which improves the performance of an anomaly detector through an iterative process. By capturing spectral and spatial information from previous iterations, ISSHAD creates a new data cube for the next iteration and uses the feedback from anomaly maps to differentiate anomalies from the background.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2023)
Article
Computer Science, Artificial Intelligence
Vitjan Zavrtanik, Matej Kristan, Danijel Skocaj
Summary: The article discusses the issue of visual anomaly detection and the shortcomings of the auto-encoder method in reconstructing anomalous regions. A new approach (RIAD) is proposed to address this problem by reconstructing images through inpainting, setting a new state-of-the-art in anomaly detection benchmarks.
PATTERN RECOGNITION
(2021)
Proceedings Paper
Computer Science, Hardware & Architecture
Noor Hadi Hammood, Bahaa Al-Musawi, Ahmed Hazim Alhilali
Summary: This article presents a survey focused on extracting all possible BGP features from a BGP control plane and determining the most significant ones for detecting BGP abnormalities. It also provides a review of recent works concerned with using Machine Learning algorithms to detect BGP anomalies.
APPLICATIONS AND TECHNIQUES IN INFORMATION SECURITY (ATIS 2021)
(2022)
Article
Computer Science, Artificial Intelligence
Talha Oktay, Erdenay Yogurtcuoglu, Ramazan Nejdet Sarikaya, Ali Recep Karaca, Mehmet Firat Komurcu, Ahmet Sayar
Summary: Logistic companies heavily rely on Vehicle Tracking Systems (VTS) to provide valuable data about vehicle condition, cargo status, and trips. However, the collected data is often underutilized due to challenges in storing, processing, analyzing, and reporting logistic data. Real-time analysis of data from VTS and mobile devices can provide significant value for businesses.
EXPERT SYSTEMS WITH APPLICATIONS
(2021)
Article
Computer Science, Hardware & Architecture
Bing Qian, Shun Lu
Summary: This paper presents a deep learning method to replace traditional expert systems with massive mobile data, demonstrating through experiments its effectiveness and superior performance in abnormality detection.
Article
Computer Science, Information Systems
Man Zeng, Dandan Li, Pei Zhang, Kun Xie, Xiaohong Huang
Summary: In this paper, a method named FL-RLD is proposed to detect route leaks while maintaining the privacy of business relationships between ASes by using a blockchain-based federated learning framework. FL-RLD performs better in detecting route leaks than the single AS detection, regardless of whether the datasets are balanced or imbalanced.
ACM TRANSACTIONS ON INTERNET TECHNOLOGY
(2023)
Article
Engineering, Electrical & Electronic
Bo Xu, Shengxin Li, Asghar A. Razzaqi, Yu Guo, Lianzhao Wang
Summary: A novel measurement information anomaly detection method based on ANFIS is proposed to accurately identify and isolate abnormal acoustic distance information, improving the stability and accuracy of the localization system.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2021)
Article
Computer Science, Artificial Intelligence
Jin Fan, Zehao Wang, Huifeng Wu, Danfeng Sun, Jia Wu, Xin Lu
Summary: This study focuses on detecting anomalies in massive volumes of multivariate time series data and proposes a new unsupervised anomaly detection model called ATF-UAD. The model reconstructs abnormal values using a time reconstructor and a frequency reconstructor, and maximizes the identification of normal and abnormal values through dual-view adversarial learning mechanism. Experimental results show an average improvement of 6.94% in terms of F1 score compared to the state-of-the-art method.
Proceedings Paper
Computer Science, Hardware & Architecture
Kevin Hoarau, Pierre Ugo Tournoux, Tahiry Razafindralambo
Summary: The Border Gateway Protocol (BGP) is responsible for route exchange at the Internet scale and anomalies can be categorized into large or small scale. Machine learning models can use either graph features or statistical features to analyze and detect anomalies in BGP behavior. Statistical features have better accuracy for large scale anomalies, while graph features increase detection accuracy by 15% for small scale anomalies and are suitable for BGP small scale anomaly detection.
PROCEEDINGS OF THE IEEE 46TH CONFERENCE ON LOCAL COMPUTER NETWORKS (LCN 2021)
(2021)
Article
Computer Science, Artificial Intelligence
Jie Yang, Yong Shi, Zhiquan Qi
Summary: In this paper, an unsupervised mechanism called deep feature correspondence (DFC) is proposed to detect and segment anomalies in images using prior knowledge from anomaly-free samples. The DFC utilizes an asymmetric dual network framework and incorporates self-feature enhancement and multi-context residual learning modules to improve detection performance. The approach achieves state-of-the-art results on the benchmark dataset and outperforms other methods in real industrial inspection scenarios.
PATTERN RECOGNITION
(2022)
Article
Computer Science, Hardware & Architecture
Feixue Han, Mowei Wang, Yong Cui, Qing Li, Ru Liang, Yashe Liu, Yong Jiang
Summary: This article focuses on the research trend of effective latency reduction designs in data center networks, particularly on reducing queuing delays in switches. It provides an overview of the three developing stages of existing schemes, discusses recent advances and their design principles, and presents the challenges and opportunities for future work.
Article
Computer Science, Information Systems
Xiaoteng Ma, Qing Li, Yong Jiang, Gabriel-Miro Muntean, Longhao Zou
Summary: Flex-Steward is a solution for multi-client joint QoE optimization during bottleneck bandwidth sharing. It reduces QoE unfairness by using an adaptive bitrate delivery algorithm based on Neural Networks and reinforcement learning.
IEEE TRANSACTIONS ON NETWORK AND SERVICE MANAGEMENT
(2022)
Article
Computer Science, Information Systems
Qing Li, Yichao Wu, Jingpu Duan, Jiahai Yang, Yong Jiang
Summary: This paper addresses the issue of the rapidly growing global routing table due to the exhaustion of IPv4 addresses and the deployment of IPv6 networks. The proposed algorithm calculates the generalized next hops of a network prefix for the aggregation of the NSFIB. Additionally, a weighted aggregation algorithm is introduced to control path stretch. Experimental results show significant reductions in FIB size and path stretch under both IPv4 and IPv6 networks.
IEEE TRANSACTIONS ON NETWORK AND SERVICE MANAGEMENT
(2022)
Article
Computer Science, Theory & Methods
Zongyi Zhao, Xingang Shi, Zhiliang Wang, Qing Li, Han Zhang, Xia Yin
Summary: HashFlow is a method for efficient and accurate collection of flow records, which tackles the challenges brought by the increase in network traffic using main flow tables and ancillary tables, and has better performance than its competitors.
IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS
(2022)
Article
Computer Science, Information Systems
Ruoyu Li, Qing Li, Jianer Zhou, Yong Jiang
Summary: Internet of Things (IoT) is undergoing rapid development and widespread deployment. However, due to the inability of low-power devices to support complex security mechanisms, they are highly vulnerable to malware attacks. This article proposes ADRIoT, an edge computing-based anomaly detection framework for IoT networks, which effectively uncovers potential threats. By utilizing LSTM autoencoders and a collaborative mechanism among multiple edge devices, ADRIoT is able to detect various IoT-based attacks and contribute to building a more secure IoT environment.
IEEE INTERNET OF THINGS JOURNAL
(2022)
Article
Engineering, Electrical & Electronic
Xiaoteng Ma, Qing Li, Longhao Zou, Junkun Peng, Jianer Zhou, Jimeng Chai, Yong Jiang, Gabriel-Miro Muntean
Summary: This paper proposes a QoE-aware Adaptive Video bitrate Aggregation scheme for HTTP live streaming based on smart edge computing (QAVA). The QAVA algorithm adapts client bitrates and caches content at the edge server to smooth live streaming traffic and improve user Quality of Experience (QoE).
IEEE TRANSACTIONS ON BROADCASTING
(2022)
Article
Engineering, Electrical & Electronic
Hanling Wang, Qing Li, Heyang Sun, Zuozhou Chen, Yingqian Hao, Junkun Peng, Zhenhui Yuan, Junsheng Fu, Yong Jiang
Summary: Edge-cloud collaborative video analytics is revolutionizing the handling, processing, and transmission of surveillance camera data worldwide. VaBUS is an innovative real-time video analytics system that utilizes contextual information to reduce bandwidth consumption for semantic compression. By sending only highly confident Region of Interests (RoIs) to the cloud through adaptive weighting and encoding, VaBUS achieves significant bandwidth reduction without sacrificing accuracy.
IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS
(2023)
Article
Mathematics, Interdisciplinary Applications
Jianhui Lv, Shen Zhao, Bo Yi, Qing Li
Summary: In the era of 5G/B5G, Mobile Edge Computing (MEC) is proposed to address the computing-intensive and delay-sensitive applications. In this paper, a multi-agent cooperation mechanism and a routing mechanism based on deep reinforcement learning (DRL) are proposed to solve the challenge of limited computing power in edge servers. The simulation results show the superiority of the proposed mechanisms in terms of network delay, throughput, task success rate, and average task response delay.
FRACTALS-COMPLEX GEOMETRY PATTERNS AND SCALING IN NATURE AND SOCIETY
(2023)
Article
Computer Science, Hardware & Architecture
Jianer Zhou, Xinyi Qiu, Zhenyu Li, Qing Li, Gareth Tyson, Jingpu Duan, Yi Wang, Heng Pan, Qinghua Wu
Summary: Most existing congestion control algorithms are designed for specific network environments and struggle to achieve universal performance. Antelope proposes a dynamic switching system for selecting the most suitable CCAs for specific flows in different environments, utilizing machine learning models and kernel-level support for dynamic adjustment. Evaluation demonstrates Antelope's effectiveness in improving throughput and reducing delay compared to traditional algorithms like BBR and CUBIC.
IEEE-ACM TRANSACTIONS ON NETWORKING
(2023)
Article
Computer Science, Hardware & Architecture
Guorui Xie, Qing Li, Guanglin Duan, Jiaye Lin, Yutao Dong, Yong Jiang, Dan Zhao, Yuan Yang
Summary: In this paper, the authors propose Mousikav2, a solution that uses binary decision trees for in-network classification in P4 switches and employs knowledge distillation to indirectly deploy complex models in switches. The approach improves classification accuracy, reduces switch stage latency, and lowers memory usage.
IEEE-ACM TRANSACTIONS ON NETWORKING
(2023)
Article
Telecommunications
Qing Li, He Huang, Yong Jiang, Jingpu Duan
Summary: The combination of network function virtualization and software-defined networking allows for processing flows based on their characteristics and needs. However, the existing NFV platform lacks a comprehensive solution for scaling under workload variation, which can negatively impact system performance. To address this issue, AdaptNF proposes a novel NFV platform that supports a combination of coarse-grained and fine-grained resource scheduling strategies. It also introduces a new algorithm to efficiently balance workload among multiple network function instances. Experimental results show that AdaptNF optimizes resource allocation and improves performance in terms of network throughput and latency.
DIGITAL COMMUNICATIONS AND NETWORKS
(2023)
Article
Computer Science, Information Systems
Qing Li, Ying Chen, Aoyang Zhang, Yong Jiang, Longhao Zou, Zhimin Xu, Gabriel-Miro Muntean
Summary: This paper proposes a flexible super-resolution-based video coding and uploading framework to improve the quality of live video streaming in limited uplink network bandwidth conditions. By employing a flexible video coding scheme and bitrate adaptation algorithm, the framework reduces the required bandwidth while maintaining video quality, thereby enhancing users' Quality of Experience.
IEEE TRANSACTIONS ON MULTIMEDIA
(2023)
Article
Computer Science, Hardware & Architecture
Chuang Hu, Rui Lu, Qianlong Sang, Huanghuang Liang, Dan Wang, Dazhao Cheng, Jin Zhang, Qing Li, Junkun Peng
Summary: This paper introduces a real-time video analytics system called Gemini, which is enhanced by a dual-image FPGA. Gemini improves analytics accuracy by pre-configuring CPU and GPU images and elastically multiplexing the dual CPU-GPU resources. The system requires hardware and software revisions, and addresses challenges through hardware abstraction and a bandit learning approach.
IEEE TRANSACTIONS ON COMPUTERS
(2023)
Proceedings Paper
Computer Science, Theory & Methods
Tao Jin, Weichao Li, Qing Li, Qianyi Huang, Yong Jiang, Shutao Xia
Summary: This paper proposes a novel Burst Queue Recovery (BQR) model for inferring the available bandwidth (ABW) of a network. By correlating the one-way delay (OWD) with the queue length variation, the model accurately calculates the ABW and is more tolerant to transient traffic bursts and supports multiple congestible links. Based on this model, the authors build a fast and accurate ABW estimation tool.
2022 IEEE/ACM 30TH INTERNATIONAL SYMPOSIUM ON QUALITY OF SERVICE (IWQOS)
(2022)
Proceedings Paper
Computer Science, Theory & Methods
Jiashuo Lin, Weichao Li, Xingbo Feng, Shuangping Zhan, Jingbin Feng, Jian Cheng, Tao Wang, Qing Li, Yi Wang, Fuliang Li, Bo Tang
Summary: This paper mathematically evaluates the performance of flow scheduling algorithms with and without network cycle, and finds that the performance can be significantly improved only when the network cycle is set to a proper value. A novel assessment metric and optimization algorithm are introduced to better evaluate the scheduling effect.
2022 IEEE/ACM 30TH INTERNATIONAL SYMPOSIUM ON QUALITY OF SERVICE (IWQOS)
(2022)
Article
Computer Science, Hardware & Architecture
Xiaolin Gu, Wenjia Wu, Yusen Zhou, Aibo Song, Ming Yang, Zhen Ling, Junzhou Luo
Summary: This study proposes a radio frequency fingerprint identification solution based on crystal oscillator temperature adjustment, which enhances the differences between Wi-Fi device fingerprints and mitigates collision. Experimental results demonstrate the effectiveness of the system in identifying smartphones under different scenarios.
Article
Computer Science, Hardware & Architecture
Yutong Wu, Jianyue Zhu, Xiao Chen, Yu Zhang, Yao Shi, Yaqin Xie
Summary: This paper proposes a quality-of-service-based SIC order method and optimizes power allocation for maximizing the rate in the uplink NOMA system. The simulation results demonstrate the superiority of the proposed method compared to traditional orthogonal multiple access and exhaustive search.
Article
Computer Science, Hardware & Architecture
Songshi Dou, Li Qi, Zehua Guo
Summary: Emerging cloud services and applications have different QoS requirements for the network. SD-WANs play a crucial role in QoS provisioning by introducing network programmability, dynamic flow routing, and low data transmission latency. However, controller failures may degrade QoS. To address this, we propose PREDATOR, a QoS-aware network programmability recovery scheme that achieves fine-grained per-flow remapping without introducing extra delays, ensuring QoS robustness for high-priority flows.
Article
Computer Science, Hardware & Architecture
Ke Wang, Xiaojuan Ma, Heng Kang, Zheng Lyu, Baorui Feng, Wenliang Lin, Zhongliang Deng, Yun Zou
Summary: This paper proposes a method based on a parallel network simulation architecture to improve the simulation efficiency of satellite networks. By effectively partitioning the network topology and using algorithms such as resource assessment and load balancing, the simulation performance is enhanced. Experimental results demonstrate the effectiveness of this method.
Article
Computer Science, Hardware & Architecture
Sijin Yang, Lei Zhuang, Julong Lan, Jianhui Zhang, Bingkui Li
Summary: This paper proposes a reuse-based online scheduling mechanism that achieves deterministic transmission of dynamic flows through dynamic path planning and coordinated scheduling of time slots. Experimental results show that the proposed mechanism improves the scheduling success rate by 37.3% and reduces time costs by up to 66.6% compared to existing online scheduling algorithms.