Federated learning has emerged as an essential technique of protecting privacy since it allows clients to train models locally without explicitly exchanging sensitive data.Extensive research has been conducted on the ...Federated learning has emerged as an essential technique of protecting privacy since it allows clients to train models locally without explicitly exchanging sensitive data.Extensive research has been conducted on the issue of data heterogeneity in federated learning,but effective model training with severely imbalanced label distributions remains an unexplored area.This paper presents a novel Cluster Federated Learning Algorithm with Intra-cluster Correction(CFIC).First,CFICselects samples fromeach cluster during each round of sampling,ensuring that no single category of data dominates the model training.Second,in addition to updating local models,CFIC adjusts its own parameters based on information shared by other clusters,allowing the final cluster models to better reflect the true nature of the entire dataset.Third,CFIC refines the cluster models into a global model,ensuring that even when label distributions are extremely imbalanced,the negative effects are significantly mitigated,thereby improving the global model’s performance.We conducted extensive experiments on seven datasets and six benchmark algorithms.Theresults show that the CFIC algorithm has a higher generalization ability than the benchmark algorithms.CFIC maintains high accuracy and rapid convergence rates even in a variety of non-independent identically distributed label skew distribution settings.The findings indicate that the proposed algorithm has the potential to become a trustworthy and practical solution for privacy preservation,whichmight be applied to fields such as medical image analysis,autonomous driving technologies,and intelligent educational platforms.展开更多
Deep reinforcement learning is broadly employed in the optimization of wireless video transmissions.Nevertheless,the instability of the deep reinforcement learning algorithm affects the further improvement of the vide...Deep reinforcement learning is broadly employed in the optimization of wireless video transmissions.Nevertheless,the instability of the deep reinforcement learning algorithm affects the further improvement of the video transmission quality.The federated learning method based on distributed data sets was used to reduce network costs and increase the learning efficiency of the deep learning network model.It solved too much data transfer costs and broke down the data silos.Intra-clustered dynamic federated deep reinforcement learning(IcD-FDRL)was constructed in clustered mobile edge-computing(CMEC)networks due to the promoted video transmission quality for the stability and efficiency of the DRL algorithm.Then,the IcD-FDRL algorithm was employed to CMEC networks’edge for intelligentedge video transmissions,which could satisfy the diversified needs of different users.The simulation analysis proved the effectiveness of IcD-FDRL in improving QoE,cache hit ratio,and training.展开更多
目的探讨基于双参数磁共振成像(biparametric magnetic resonance imaging,bpMRI)生境成像分析诊断高风险和低风险前列腺癌(prostate cancer,PCa)的价值。材料与方法回顾性分析了2023年12月至2024年8月期间在马鞍山人民医院就诊期间,经...目的探讨基于双参数磁共振成像(biparametric magnetic resonance imaging,bpMRI)生境成像分析诊断高风险和低风险前列腺癌(prostate cancer,PCa)的价值。材料与方法回顾性分析了2023年12月至2024年8月期间在马鞍山人民医院就诊期间,经穿刺或手术病理证实为PCa患者191例,其中高风险PCa为131例,低风险PCa为60例。将191例患者按7∶3的比例随机分为训练集及测试集,所有患者均进行bpMRI扫描,对T2WI、小视野扩散加权成像(ZOOMit diffusion weighted imaging,ZOOMit-DWI)、表观扩散系数(apparent diffusion coefficient,ADC)序列进行预处理。在ADC图像中提取19个影像组学特征,通过整合T2WI与ZOOMit-DWI图像,采用无监督K均值聚类方法在所有肿瘤体素中生成相似亚区。根据生境亚区结果计算出191例患者的肿瘤内异质性分数(intratumoral heterogeneity score,ITHscore),从亚区提取组学特征,对影像组学特征进行降维及过滤,筛选出相关系数最高的组学特征,通过SHAP分析可视化图,以评估各特征重要性。经过各特征融合和筛选,分别建立10个生境影像组学模型。计算每个模型的阈值、敏感度、特异度、准确度、阴性预测值和阳性预测值,绘制受试者工作特征(receiver operating characteristic,ROC)曲线并计算曲线下面积(area under the curve,AUC)分析各个模型的诊断效能,进一步使用决策曲线分析(decision curve analysis,DCA)评估模型的净获益。结果在训练集及测试集中总前列腺特异性抗原(total prostate specific antigen,tPSA)在高风险及低风险PCa中的差异有统计学意义(P<0.05)。根据Calinski-Harabasz(CH)值,确定2作为最佳生境亚区数量,通过SHAP分析,其中h2亚区中original_glszm_ZoneEntropy特征影响力最大。采用10个分类器进行模型构建,对比生境影像组学模型与临床+前列腺影像报告和数据系统(prostate imaging reporting and data system,PI-RADS)评分模型,生境影像组学模型在测试集中极端随机树(extremely randomized trees,ExtraTrees)模型预测效果最佳,训练集AUC值为0.838[95%置信区间(confidence interval,CI):0.768~0.908],测试集AUC值为0.796(95%CI:0.665~0.927)。临床数据和PI-RADS评分模型构建,逻辑回归(logistic regression,LR)模型预测效能最佳,其中训练集AUC值为0.786(95%CI:0.705~0.866),测试集AUC值为0.719(95%CI:0.550~0.887),利用DCA评估两者的临床净收益,生境影像组学模型AUC值高于临床+PI-RADS评分模型。结论基于bpMRI生境成像分析能较准确地诊断PCa的病理分型,有助于PCa的临床诊断及风险预测。展开更多
提出了一种基于簇结构的无线传感器网络数据收集协议EADEEG(an energy-aware data gathering protocol for wireless sensor networks).EADEEG通过最小化网络通信开销以及良好的能量负载平衡方法,可以有效地延长网络寿命.与以前的相关...提出了一种基于簇结构的无线传感器网络数据收集协议EADEEG(an energy-aware data gathering protocol for wireless sensor networks).EADEEG通过最小化网络通信开销以及良好的能量负载平衡方法,可以有效地延长网络寿命.与以前的相关研究相比,EADEEG采用了一种全新的簇头竞争参数,能够更好地解决节点能量异构问题.此外,EADEEG也采用了一种简单而有效的簇内节点调度算法,通过控制活动节点的密度,可以在不增加额外控制开销的条件下关闭冗余节点并保证覆盖要求,因此可以进一步延长网络寿命.模拟实验证明,在节点初始能量同构和异构两种情况下,EADEEG协议都能够满足用户对覆盖率的要求,并在网络寿命上大幅度优于LEACH(low energy adaptive clustering hierarchy),PEGASIS(power-efficient gathering in sensor information systems)和DEEG(distributed energy-efficient data gathering and aggregation protocol)协议.展开更多
In wireless sensor networks（WSNs）,appropriate topology control（TC）could efficiently balance the load among sensor nodes and extend network lifespan.Clustering is an effective topology control technique that could ...In wireless sensor networks（WSNs）,appropriate topology control（TC）could efficiently balance the load among sensor nodes and extend network lifespan.Clustering is an effective topology control technique that could reduce energy consumption and provide scalability to WSNs.However,some clustering algorithms,including the traditional low energy adaptive clustering hierarchy（LEACH）,don＇t consider the residual energy and the communication distance.The energy consumption could dramatically increase in the case of long communication distance and high rate of control message exchange.In this paper we propose an energy-balanced clustering algorithm which considers the communication distance and the residual energy.Moreover the cluster head（CH）reselection is relevant to the current CH residual energy in order to reduce overheads.The simulation results demonstrate that the proposed algorithm prolongs the lifetime of the WSN in comparison to the LEACH and a hybrid clustering approach（HCA）.展开更多
基金supported by National Natural Science Foundation of China under Grant(No.62277043)Science and Technology Research Project of Chongqing Education Commission under Grant(No.KJZD-K202300515).
文摘Federated learning has emerged as an essential technique of protecting privacy since it allows clients to train models locally without explicitly exchanging sensitive data.Extensive research has been conducted on the issue of data heterogeneity in federated learning,but effective model training with severely imbalanced label distributions remains an unexplored area.This paper presents a novel Cluster Federated Learning Algorithm with Intra-cluster Correction(CFIC).First,CFICselects samples fromeach cluster during each round of sampling,ensuring that no single category of data dominates the model training.Second,in addition to updating local models,CFIC adjusts its own parameters based on information shared by other clusters,allowing the final cluster models to better reflect the true nature of the entire dataset.Third,CFIC refines the cluster models into a global model,ensuring that even when label distributions are extremely imbalanced,the negative effects are significantly mitigated,thereby improving the global model’s performance.We conducted extensive experiments on seven datasets and six benchmark algorithms.Theresults show that the CFIC algorithm has a higher generalization ability than the benchmark algorithms.CFIC maintains high accuracy and rapid convergence rates even in a variety of non-independent identically distributed label skew distribution settings.The findings indicate that the proposed algorithm has the potential to become a trustworthy and practical solution for privacy preservation,whichmight be applied to fields such as medical image analysis,autonomous driving technologies,and intelligent educational platforms.
基金supported by the Start-up Project of Doctoral Research in Jiangxi University of Water Resources and Electric Power(No.2024kyqd062)the Key Project of Science and Technology Research of Jiangxi Provincial Education Department(No.GJJ180251)the National Natural Science Foundation of China(No.61961021).
文摘Deep reinforcement learning is broadly employed in the optimization of wireless video transmissions.Nevertheless,the instability of the deep reinforcement learning algorithm affects the further improvement of the video transmission quality.The federated learning method based on distributed data sets was used to reduce network costs and increase the learning efficiency of the deep learning network model.It solved too much data transfer costs and broke down the data silos.Intra-clustered dynamic federated deep reinforcement learning(IcD-FDRL)was constructed in clustered mobile edge-computing(CMEC)networks due to the promoted video transmission quality for the stability and efficiency of the DRL algorithm.Then,the IcD-FDRL algorithm was employed to CMEC networks’edge for intelligentedge video transmissions,which could satisfy the diversified needs of different users.The simulation analysis proved the effectiveness of IcD-FDRL in improving QoE,cache hit ratio,and training.
基金Supported by the National Natural Science Foundation of China under Grant No.60573132(国家自然科学基金)the National Grand Fundamental Research973Program of China under Grant No.2006CB303000(国家重点基础研究发展规划(973))the Hong Kong Polytechnic University under Grant No.A-PF77(香港理工大学)
文摘提出了一种基于簇结构的无线传感器网络数据收集协议EADEEG(an energy-aware data gathering protocol for wireless sensor networks).EADEEG通过最小化网络通信开销以及良好的能量负载平衡方法,可以有效地延长网络寿命.与以前的相关研究相比,EADEEG采用了一种全新的簇头竞争参数,能够更好地解决节点能量异构问题.此外,EADEEG也采用了一种简单而有效的簇内节点调度算法,通过控制活动节点的密度,可以在不增加额外控制开销的条件下关闭冗余节点并保证覆盖要求,因此可以进一步延长网络寿命.模拟实验证明,在节点初始能量同构和异构两种情况下,EADEEG协议都能够满足用户对覆盖率的要求,并在网络寿命上大幅度优于LEACH(low energy adaptive clustering hierarchy),PEGASIS(power-efficient gathering in sensor information systems)和DEEG(distributed energy-efficient data gathering and aggregation protocol)协议.
基金Supported by the National Natural Science Foundation of China(6104086)Scientific Research,Postgraduate Training Joint-Build Project(20120639002)
文摘In wireless sensor networks（WSNs）,appropriate topology control（TC）could efficiently balance the load among sensor nodes and extend network lifespan.Clustering is an effective topology control technique that could reduce energy consumption and provide scalability to WSNs.However,some clustering algorithms,including the traditional low energy adaptive clustering hierarchy（LEACH）,don＇t consider the residual energy and the communication distance.The energy consumption could dramatically increase in the case of long communication distance and high rate of control message exchange.In this paper we propose an energy-balanced clustering algorithm which considers the communication distance and the residual energy.Moreover the cluster head（CH）reselection is relevant to the current CH residual energy in order to reduce overheads.The simulation results demonstrate that the proposed algorithm prolongs the lifetime of the WSN in comparison to the LEACH and a hybrid clustering approach（HCA）.
