检索结果-内蒙古大学图书馆

Novel kernel density estimator based on ensemble unbiased cross-validation

INFORMATION SCIENCES 2021年 581卷 327-344页

作者： He, Yu-Lin Ye, Xuan Huang, De-Fa Huang, Joshua Zhexue Zhai, Jun-Hai Shenzhen Univ Big Data Inst Coll Comp Sci & Software Engn Shenzhen 518060 Peoples R China Shenzhen Univ Natl Engn Lab Big Data Syst Comp Technol Shenzhen 518060 Peoples R China Hebei Univ Coll Math & Informat Sci Hebei Key Lab Machine Learning & Computat Intelli Baoding 071002 Peoples R China

Unbiased cross-validation (UCV) is a commonly-used method to calculate the optimal bandwidth for the kernel density estimator (KDE), which estimates the underlying probability density function (PDF) for a given data set. Since the UCV method was proposed, there have been few studies that have pointed out its instability when determining the KDE bandwidth. Following the principle of stability improvement, this paper presents a novel ensemble UCV based KDE (EUCV-KDE), which determines the expectation of an estimated PDF using an ensemble of data-block based UCVs rather than a single data-point based UCV. To derive the optimal bandwidth, a novel objective function is designed for EUCV-KDE by considering the empirical and structural risk of KDE together. We validate the rationality and effectiveness of EUCV-KDE on 10 probability distributions. The experimental results show that EUCV-KDE is convergent as the number of data-block based UCVs increases and can obtain a more stable and better prediction performance than the classical UCV-KDE and the revisited cross-validation (RCV) based KDE (RCV-KDE). In addition, a real-world application based on UK climate data is provided to further validate the effectiveness of EUCV-KDE by determining the optimal bandwidth for Nadaraya-Watson kernel regression estimator. (c) 2021 Elsevier Inc. All rights reserved.

关键词： Probability density function Kernel density estimator Bandwidth Unbiased cross-validation data-block based UCV data-point based UCV

来源：评论

学校读者我要写书评

暂无评论

Towards Efficient Cross-Modal Anomaly Detection Using Triple-Adaptive Network and Bi-Quintuple Contrastive learning

引用

IEEE TRANSACTIONS ON EMERGING TOPICS IN COMPUTATIONAL INTELLIGENCE 2024年第1期8卷 697-709页

作者： Peng, Shu-Juan Fan, Ye Cheung, Yiu-ming Liu, Xin Cui, Zhen Li, Taihao Huaqiao Univ Dept Artificial Intelligence Xiamen 361021 Peoples R China Artificial Intelligence Res Inst Zhejiang Lab Hangzhou 311121 Peoples R China Fujian Prov Univ Huaqiao Univ Key Lab Comp Vis & Machine Learning Xiamen 361021 Peoples R China Xiamen Key Lab Comp Vis & Pattern Recognit Xiamen 361021 Peoples R China Huaqiao Univ Fujian Key Lab Big Data Intelligence & Secur Xiamen 361021 Peoples R China Hong Kong Baptist Univ Dept Comp Sci Hong Kong Peoples R China Nanjing Univ Sci & Technol Key Lab Intelligent Percept & Syst High Dimens Inf Minist Educ Nanjing 210094 Peoples R China

Cross-modal anomaly detection is a relatively new and challenging research topic in machine learning field, which aims at identifying the anomalies whose patterns are disparate across different modalities. As far as we know, this topic has yet to be well studied, and existing works often suffer from the incomplete anomalous data detection and low data utilization problems. To alleviate these limitations, this paper proposes an efficient deep cross-modal anomaly detection approach via Triple-adaptive Network and Bi-quintuple Contrastive learning (TN-BCL), which lies among the earliest attempt to detect various cross-modal anomalies within the heterogeneous multi-modal data. To be specific, a triple-adaptive network is explicitly designed to identify various anomalies, whose patterns are disparate in both single-modal scenario and cross-modal scenario. On the one hand, the top branch network is utilized to adaptively detect the attribute anomalies and part of mixed anomalies in multi-modal data samples. On the other hand, the bottom two-branch network, with shared residual blocks, is leveraged to learn the discriminative cross-modal embeddings. At the same time, an efficient bi-quintuple contrastive learning method is designed to enhance the feature correlation between the same attribute data, while maximally enlarging the feature difference between different attribute data. Besides that, the bidirectional learning scheme is employed to significantly improve the data utilization. Through the joint exploitation of the above, different kinds of anomalous samples can be well detected across different modalities. Extensive experiments show that the proposed framework outperforms the state-of-the-art competing methods, with a large improvement margin.

关键词： Anomaly detection Detectors Behavioral sciences Feature extraction Task analysis Electronic mail Computer science Cross-modal anomaly detection triple-adaptive network bi-quintuple contrastive learning shared residual block

来源：评论

学校读者我要写书评

暂无评论

MSI-UNET: A FLEXIBLE UNET-BASED MULTI-SCALE INTERACTIVE FRAMEWORK FOR 3D GASTRIC TUMOR SEGMENTATION ON CT SCANS 21

MSI-UNET: A FLEXIBLE UNET-BASED MULTI-SCALE INTERACTIVE FRAM...

引用

21st IEEE International Symposium on Biomedical Imaging (ISBI)

作者： Chen, Heyun Chen, Zifan Zhao, Jie Li, Haoshen Li, Jiazheng Liu, Yiting Yuan, Mingze Bao, Peng Nan, Xinyu Dong, Bin Tang, Lei Zhang, Li Peking Univ Ctr Data Sci Beijing Peoples R China Peking Univ Natl Engn Lab Big Data Anal & Applicat Beijing Peoples R China Peking Univ Canc Hosp & Inst Beijing Peoples R China Peking Univ Beijing Int Ctr Math Res Beijing Peoples R China Peking Univ Ctr Machine Learning Res Beijing Peoples R China Peking Univ Changsha Inst Comp & Digital Econ Beijing Peoples R China

ISBN: (纸本)9798350313345;9798350313338

Accurate segmentation of gastric tumors is critical yet presents a formidable challenge in medical imaging, where conventional UNet-based frameworks, despite their prevalence, falter on intricate tumor samples due to their limited interactive capacities. The SAM-based segmentation methods address this shortcoming yet with insufficient accuracy. By ingeniously blending images with mask inputs, our MSI-UNet leverages a U-shaped design to deliver pixel-perfect segmentation accuracy, while a novel multi-scale attention module adeptly harnesses interaction points for refined information extraction. When benchmarked on gastric tumor segmentation tasks, MSI-UNet surpasses existing state-of-the-art methods, elevating the Dice Similarity Coefficient (DSC) from 74.82% to 79.3% and minimizing Average Surface Distance (ASD) from 6.46 to 1.98, achieving a comparable accuracy with inter-radiologist consistency of 79.7% DSC. Furthermore, our framework demonstrates superior predictive prowess in survival analysis, enhancing the C-index from 61.7% to 68.68%. Ample experimental comparisons have substantiated that MSI-UNet holds the potential to offer considerable assistance to healthcare professionals in managing and decoding subsequent medical procedures.

关键词： Gastric tumor segmentation interactive segmentation deep learning

来源：评论

学校读者我要写书评

暂无评论

FedDP-SA: Boosting Differentially Private Federated learning via Local data Set Splitting

引用

IEEE INTERNET OF THINGS JOURNAL 2024年第19期11卷 31687-31698页

作者： Liu, Xuezheng Zhou, Yipeng Wu, Di Hu, Miao Hui Wang, Jessie Guizani, Mohsen Sun Yat Sen Univ Sch Comp Sci & Engn Guangdong Key Lab Big Data Anal & Proc Guangzhou 510006 Peoples R China Macquarie Univ Fac Sci & Engn Dept Comp Sydney NSW 2109 Australia Tsinghua Univ Inst Network Sci & Cyberspace Beijing 100084 Peoples R China Tsinghua Univ Beijing Natl Res Ctr Informat Sci & Technol Beijing 100084 Peoples R China Mohamed bin Zayed Univ Artificial Intelligence Machine Learning Dept Abu Dhabi U Arab Emirates

Federated learning (FL) emerges as an attractive collaborative machine learning framework that enables training of models across decentralized devices by merely exposing model parameters. However, malicious attackers can still hijack communicated parameters to expose clients' raw samples resulting in privacy leakage. To defend against such attacks, differentially private FL (DPFL) is devised, which incurs negligible computation overhead in protecting privacy by adding noises. Nevertheless, the low model utility and communication efficiency makes DPFL hard to be deployed in the real environment. To overcome these deficiencies, we propose a novel DPFL algorithm called FedDP-SA (namely, federated learning with differential privacy by splitting Local data sets and averaging parameters). Specifically, FedDP-SA splits a local data set into multiple subsets for parameter updating. Then, parameters averaged over all subsets plus differential privacy (DP) noises are returned to the parameter server. FedDP-SA offers dual benefits: 1) enhancing model accuracy by efficiently lowering sensitivity, thereby reducing noise to ensure DP and 2) improving communication efficiency by communicating model parameters with a lower frequency. These advantages are validated through sensitivity analysis and convergence rate analysis. Finally, we conduct comprehensive experiments to verify the performance of FedDP-SA compared with other state-of-the-art baseline algorithms.

关键词： Noise Privacy Computational modeling data models Differential privacy Accuracy Internet of Things data splitting federated learning (FL) Gaussian mechanism sensitivity and convergence rate

来源：评论

学校读者我要写书评

暂无评论

Computationally Efficient Approximations for Matrix-Based Renyi's Entropy

引用

IEEE TRANSACTIONS ON SIGNAL PROCESSING 2022年 70卷 6170-6184页

作者： Gong, Tieliang Dong, Yuxin Yu, Shujian Dong, Bo Jiaotong Univ Sch Comp Sci & Technol Xian 710049 Peoples R China Shaanxi Prov Key Lab Big Data Knowledge Engn Xian 710049 Peoples R China UiT Arctic Univ Norway Machine Learning Grp N-9019 Tromso Norway Vrije Univ Amsterdam Dept Comp Sci NL-1081 HV Amsterdam Netherlands

The recently developed matrix-based Renyi's alpha-order entropy enables measurement of information in data simply using the eigenspectrum of symmetric positive semi-definite (PSD) matrices in reproducing kernel Hilbert space, without estimation of the underlying data distribution. This intriguing property makes this new information measurement widely adopted in multiple statistical inference and learning tasks. However, the computation of such quantity involves the trace operator on a PSD matrix G to power alpha (i.e., tr(G(alpha))), with a normal complexity of nearly O (n(3)), which severely hampers its practical usage when the number of samples (i.e., n) is large. In this work, we present computationally efficient approximations to this new entropy functional that can reduce its complexity to even significantly less than O (n(2)). To this end, we leverage the recent progress on Randomized Numerical Linear Algebra, developing Taylor, Chebyshev and Lanczos approximations to tr(G(alpha) ) for arbitrary values of alpha by converting it into a matrix-vector multiplication problem. We also establish the connection between the matrix-based Renyi's entropy and PSD matrix approximation, which enables exploiting both clustering and block low-rank structure of G to further reduce the computational cost. We theoretically provide approximation accuracy guarantees and illustrate the properties for different approximations. Large-scale experimental evaluations on both synthetic and real-world data corroborate our theoretical findings, showing promising speedup with negligible loss in accuracy.

关键词： Entropy Kernel Estimation Symmetric matrices Signal processing algorithms Eigenvalues and eigenfunctions Approximation algorithms Matrix-based Renyi's entropy randomized numerical linear algebra trace estimation information bottleneck mutual information

来源：评论

学校读者我要写书评

暂无评论

Modified metaheuristics with stacked sparse denoising autoencoder model for cervical cancer classification

引用

COMPUTERS & ELECTRICAL ENGINEERING 2022年 103卷

作者： Vaiyapuri, Thavavel Alaskar, Haya Syed, Liyakathunisa Aljohani, Eman Alkhayyat, Ahmed Shankar, K. Kumar, Sachin Prince Sattam Bin Abdulaziz Univ Coll Comp Engn & Sci Dept Comp Sci Al Kharj Saudi Arabia Taibah Univ Coll Comp Sci & Engn Dept Comp Sci Madinah Saudi Arabia Islamic Univ Coll Tech Engn Najaf Iraq South Ural State Univ Big Data & Machine Learning Lab Chelyabinsk 454080 Russia

Cervical cancer is the most commonly diagnosed cancer among women globally, with high mortality rate. For early diagnosis, automated and accurate cervical cancer classification ap-proaches can be developed through effective classification of Pap smear cell images. The current study introduces a novel Modified Firefly Optimization Algorithm with Deep learning-enabled cervical cancer classification (MFFOA-DL3) model for the classification of Pap Smear Images (PSI). The proposed MFFOA-DL3 model examines the PSI for the existence of cervical cancer cells. To accomplish this, the proposed MFFOA-DL3 model primarily applies Bilateral Filtering (BF) -based noise removal approach to get rid of the noise. Then, Kapur's entropy-based image seg-mentation technique is applied to determine the affected regions. Moreover, EfficientNet tech-nique is also applied to generate the feature vectors. Finally, MFFOA with Stacked Sparse Denoising Autoencoder (SSDA) model is exploited to classify the PSI. In current study, MFFOA is utilized to appropriately modify the parameters related to SSDA model. The proposed MFFOA-DL3 model was experimentally validated using benchmark dataset. The results attained from extensive comparative analysis highlighted the better performance of MFFOA-DL3 model over other recent approaches.

关键词： Pap smear images Computer aided diagnosis Medical imaging Cervical cancer Deep neural network Image-guided intervention Computer vision Decision making

来源：评论

学校读者我要写书评

暂无评论

Spatial transition tensor of single cells

引用

NATURE METHODS 2024年第6期21卷 1053-1062页

作者： Zhou, Peijie Bocci, Federico Li, Tiejun Nie, Qing Univ Calif Irvine Dept Math Irvine CA 92697 USA Peking Univ LMAM Beijing Peoples R China Peking Univ Sch Math Sci Beijing Peoples R China Univ Calif Irvine Dept Cell & Dev Biol Irvine CA 92697 USA Peking Univ Ctr Machine Learning Res Beijing Peoples R China AI Sci Inst Beijing Peoples R China Natl Engn Lab Big Data Anal & Applicat Beijing Peoples R China

Spatial transcriptomics and messenger RNA splicing encode extensive spatiotemporal information for cell states and transitions. The current lineage-inference methods either lack spatial dynamics for state transition or cannot capture different dynamics associated with multiple cell states and transition paths. Here we present spatial transition tensor (STT), a method that uses messenger RNA splicing and spatial transcriptomes through a multiscale dynamical model to characterize multistability in space. By learning a four-dimensional transition tensor and spatial-constrained random walk, STT reconstructs cell-state-specific dynamics and spatial state transitions via both short-time local tensor streamlines between cells and long-time transition paths among attractors. Benchmarking and applications of STT on several transcriptome datasets via multiple technologies on epithelial-mesenchymal transitions, blood development, spatially resolved mouse brain and chicken heart development, indicate STT's capability in recovering cell-state-specific dynamics and their associated genes not seen using existing methods. Overall, STT provides a consistent multiscale description of single-cell transcriptome data across multiple spatiotemporal scales. STT is a method that connects mRNA splicing and cell-state transitions across spatiotemporal scales and at single-cell resolution.

关键词： Computational biophysics Computational models Gene regulatory networks

来源：评论

学校读者我要写书评

暂无评论

CPFedAvg: Enhancing Hierarchical Federated learning via Optimized Local Aggregation and Parameter Mixing

IEEE TRANSACTIONS ON NETWORKING

引用

IEEE TRANSACTIONS ON NETWORKING 2025年第3期33卷 1160-1173页

作者： Liu, Xuezheng Zhou, Yipeng Wu, Di Hu, Miao Chen, Min Guizani, Mohsen Sheng, Quan Z. Sun Yat sen Univ Sch Comp Sci & Engn Guangzhou 510006 Guangdong Peoples R China Guangdong Key Lab Big Data Anal & Proc Guangzhou 510006 Peoples R China Macquarie Univ Fac Sci & Engn Sch Comp Sydney NSW 2109 Australia South China Univ Technol Sch Comp Sci & Engn Guangzhou 510640 Peoples R China Pazhou Lab Guangzhou 510640 Peoples R China Mohamed BinZayed Univ Artificial Intelligence MBZU Machine Learning Dept Abu Dhabi U Arab Emirates

Hierarchical federated learning (HFL) improves the scalability and efficiency of traditional federated learning (FL) by incorporating a hierarchical topology into the FL framework. In a typical HFL system, clients are divided into multiple tiers, and the training process involves both local and global model aggregation. However, existing HFL approaches have several significant drawbacks. Firstly, the root parameter server (PS) is vulnerable to single-point failure and also acts as a bottleneck for global aggregation. Additionally, frequent global aggregation over the wide area network (WAN) incurs substantial communication costs, which negatively affect training efficiency. In this paper, we propose a novel HFL algorithm called CPFedAvg to address the aforementioned challenges. CPFedAvg introduces a root-free hierarchical topology, where the top tier consists of multiple PSes, effectively resolving the issues associated with the root PS. Additionally, we substitute the expensive global aggregation with parameter mixing operations between the PSes in the top tier. We analyze the convergence rate of CPFedAvg under non-convex loss. Based on this analysis, we formulate a convex optimization problem to optimize the frequency of executing local aggregations between consecutive parameter mixing operations. To simulate real-world communication networks, we develop FedNetSimulator to simulate a diverse range of FL communication processes. Finally, we conduct extensive experiments using real datasets (i.e., CIFAR-10 and CIFAR-100). The experimental results demonstrate that CPFedAvg can improve model accuracy by up to 18% and the speedup can be as high as 6 compared with the state-of-the-art baselines.

关键词： Topology Convergence Training Servers Federated learning Costs Wide area networks Scalability data models Clustering algorithms Parallel federated learning topology-friendly approach convergence rate optimal mixing interval

来源：评论

学校读者我要写书评

暂无评论

A Comprehensive Detection Method for the Lateral Movement Stage of APT Attacks

引用

IEEE INTERNET OF THINGS JOURNAL 2024年第5期11卷 8440-8447页

作者： He, Daojing Gu, Hongjie Zhu, Shanshan Chan, Sammy Guizani, Mohsen Guizhou Univ State Key Lab Publ Big Data Guiyang 550025 Guizhou Peoples R China Harbin Inst Technol Sch Comp Sci & Technol Shenzhen 518055 Peoples R China East China Normal Univ Sch Software Engn Shanghai 200241 Peoples R China Harbin Inst Technol Sch Econ & Management Shenzhen 518055 Peoples R China City Univ Hong Kong Dept Elect Engn Hong Kong Peoples R China Mohamed Bin Zayed Univ Artificial Intelligence Machine Learning Dept Abu Dhabi U Arab Emirates

Due to the outbreak of the new crown epidemic, more companies prefer to use telecommuting for work, which also provides more attack surfaces for APT attacks. After initially gaining access to the intranet, attackers will use server message block (SMB), RDP, and other remote sharing or connection protocols to move horizontally to achieve the purpose of privilege escalation. In this work, we design a multidimensional detection framework to detect lateral movement behavior based on the SMB protocol in the intranet environment. This framework combines active trapping and passive scanning, and uses neural networks to determine the attack samples used by the adversary when moving laterally. We test the effectiveness of the active trapping technology in a simulation environment, and verify through real malware samples that the accuracy of neural network detection can reach about 90%. The experimental results show that our work can effectively detect the lateral movement behavior using the SMB protocol in the intranet environment.

关键词： Behavioral sciences Protocols Servers Security Internet of Things Telecommunication traffic Complex networks Advanced persistent threats (APTs) computer network security

来源：评论

学校读者我要写书评

暂无评论

HC-TUS: Human Cognition-Based Trust Update Scheme for AI-Enabled VANET

引用

IEEE NETWORK 2023年第5期37卷 247-252页

作者： Fang, Weidong Zhu, Chunsheng Guizani, Mohsen Rodrigues, Joel J. P. C. Zhang, Wuxiong Chinese Acad Sci Sci & Technol Microsyst Lab Shanghai Inst Microsyst & Informat Technol Shanghai Peoples R China Univ Chinese Acad Sci Beijing Peoples R China Shanghai Res & Dev Ctr Micronano Elect Shanghai Peoples R China Shenzhen Technol Univ Coll Big Data & Internet Shenzhen Peoples R China Mohamed Bin Zayed Univ Artificial Intelligence MBZ Machine Learning Dept Abu Dhabi U Arab Emirates Lusofona Univ COPELABS Lisbon Portugal

Recently, Artificial Intelligence (AI) has received more attention for being used in many applications. It is expected to play a key role in Vehicular Ad Hoc Networks (VANET). On the other hand, AI-enabled VANET (AI-VANET) has become an emerging field. However, its cyber security is facing enormous challenges. Although many trust schemes are proposed for addressing these issues, the reliance on only trust updates could increase the risk of long-term attacks before being detected. In this article, we design a human cognition-based trust update scheme (HC-TUS) for AI-VANET. Significantly, the novel trust update scheme is designed by strategically incorporating the Ebbinghaus forgetting theory. The simulation results indicate that 1) HC-TUS could better meet the principle of & Ograve;Hard to get, easy to lose & Oacute;for trust than BRSN and BTDS;2) HC-TUS could detect and resist the collusion attack more quickly than BRSN and BTDS. The open issues in terms of trust for AI-VANET are also investigated and highlighted.

关键词： Vehicular ad hoc networks Artificial intelligence Computational modeling Security Trust management Blockchains Cognition Human factors Simulation Computer crime Artificial Intelligence (AI) Vehicular Ad Hoc Network (VANET) Trust Model Collusion Attack

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：