随着物联网(Internet of Things,IoT)设备的普及,使用入侵检测来保护IoT设备免受恶意攻击至关重要。但是,IoT的数据稀缺性限制了传统入侵检测方法的效果。同时,现有基于域自适应的入侵检测方法的对齐方式粗糙,忽略了内在语义属性的转移...随着物联网(Internet of Things,IoT)设备的普及,使用入侵检测来保护IoT设备免受恶意攻击至关重要。但是,IoT的数据稀缺性限制了传统入侵检测方法的效果。同时,现有基于域自适应的入侵检测方法的对齐方式粗糙,忽略了内在语义属性的转移,降低了特征的可区分性。为解决上述问题,提出了一种基于Transformer的域自适应物联网入侵检测(Transformer-Based Domain-Adaptive IoT Intrusion Detection,TDAIID)模型,从域间、类间和样本间3个层次对齐互联网入侵(Network Intrusion,NI)域和物联网入侵(Internet of Things Intrusion,II)域。交叉注意力机制聚焦于NI源域和II目标域中相同类别样本之间的相似特征,实现样本级别的域特征对齐;多重几何语义对齐从域级和类级两个角度进行语义对齐,有助于交叉注意力机制学习更丰富、更准确的源NI域知识。此外,为了充分挖掘未标记II目标域的潜力,从几何角度提出了一种动态中心感知伪标签算法,用于提高伪标签标记的准确性,有效降低错误分配伪标签造成的负迁移。在多个常用入侵检测数据集上的综合实验表明,TDAIID模型的性能优于当前先进的基线模型。展开更多
In the traditional environment, the factors for considering the location of the waste transfer station and the landfill are relatively fixed, and the scale of the problem is small. But in Internet of Things(IoT) envir...In the traditional environment, the factors for considering the location of the waste transfer station and the landfill are relatively fixed, and the scale of the problem is small. But in Internet of Things(IoT) environment, the waste storage in the household waste can be monitored in real time, the environmental data can be collected by means of emerging information technology, and the residents are more sensitive to the environmental pollution of the waste. Under such conditions, the method for location of traditional waste disposal facilities needs to be redeveloped to obtain a waste transfer station and landfill site that are suitable for the IoT environment. For this reason, a two-objective integer programming model is designed. The two objectives are lowest cost and minimum impact of waste on residents. The expectations of city managers and residents are considered into the modeling. Through the simulation experiments on different scale problems, the integration method for integer programming model and simulation system is verified to solve the location of waste transfer stations in IoT environment.展开更多
为了解决联邦学习在车联网中终端设备数据的异质性导致模型训练准确率不稳定和性能下降,以及车辆分布广泛,通信和计算资源有限的问题,提出一种数据类型和数据规模并行优化的群联邦迁移学习数据共享方法(swarm federated transfer learni...为了解决联邦学习在车联网中终端设备数据的异质性导致模型训练准确率不稳定和性能下降,以及车辆分布广泛,通信和计算资源有限的问题,提出一种数据类型和数据规模并行优化的群联邦迁移学习数据共享方法(swarm federated transfer learning,SFTL)。提出基于高斯混合模型的共识设备组划分机制,通过对数据分布建模构建共识设备组,实现对异质性数据的有效管理和分析;面向划分的共识设备组,设计蜂群学习训练机制,加强相似设备组之间的协同学习过程;提出组间迁移学习机制,通过模型预训练法增量迁移不同共识设备组信息最小化模型差异,提高联邦模型聚合准确率。在公共数据集上的实验表明:与基线方法相比,SFTL模型训练准确率平均提高7%,通信时间平均降低10%。展开更多
文摘随着物联网(Internet of Things,IoT)设备的普及,使用入侵检测来保护IoT设备免受恶意攻击至关重要。但是,IoT的数据稀缺性限制了传统入侵检测方法的效果。同时,现有基于域自适应的入侵检测方法的对齐方式粗糙,忽略了内在语义属性的转移,降低了特征的可区分性。为解决上述问题,提出了一种基于Transformer的域自适应物联网入侵检测(Transformer-Based Domain-Adaptive IoT Intrusion Detection,TDAIID)模型,从域间、类间和样本间3个层次对齐互联网入侵(Network Intrusion,NI)域和物联网入侵(Internet of Things Intrusion,II)域。交叉注意力机制聚焦于NI源域和II目标域中相同类别样本之间的相似特征,实现样本级别的域特征对齐;多重几何语义对齐从域级和类级两个角度进行语义对齐,有助于交叉注意力机制学习更丰富、更准确的源NI域知识。此外,为了充分挖掘未标记II目标域的潜力,从几何角度提出了一种动态中心感知伪标签算法,用于提高伪标签标记的准确性,有效降低错误分配伪标签造成的负迁移。在多个常用入侵检测数据集上的综合实验表明,TDAIID模型的性能优于当前先进的基线模型。
基金Supported by the National Natural Science Foundation of China(71531009).
文摘In the traditional environment, the factors for considering the location of the waste transfer station and the landfill are relatively fixed, and the scale of the problem is small. But in Internet of Things(IoT) environment, the waste storage in the household waste can be monitored in real time, the environmental data can be collected by means of emerging information technology, and the residents are more sensitive to the environmental pollution of the waste. Under such conditions, the method for location of traditional waste disposal facilities needs to be redeveloped to obtain a waste transfer station and landfill site that are suitable for the IoT environment. For this reason, a two-objective integer programming model is designed. The two objectives are lowest cost and minimum impact of waste on residents. The expectations of city managers and residents are considered into the modeling. Through the simulation experiments on different scale problems, the integration method for integer programming model and simulation system is verified to solve the location of waste transfer stations in IoT environment.
文摘为了解决联邦学习在车联网中终端设备数据的异质性导致模型训练准确率不稳定和性能下降,以及车辆分布广泛,通信和计算资源有限的问题,提出一种数据类型和数据规模并行优化的群联邦迁移学习数据共享方法(swarm federated transfer learning,SFTL)。提出基于高斯混合模型的共识设备组划分机制,通过对数据分布建模构建共识设备组,实现对异质性数据的有效管理和分析;面向划分的共识设备组,设计蜂群学习训练机制,加强相似设备组之间的协同学习过程;提出组间迁移学习机制,通过模型预训练法增量迁移不同共识设备组信息最小化模型差异,提高联邦模型聚合准确率。在公共数据集上的实验表明:与基线方法相比,SFTL模型训练准确率平均提高7%,通信时间平均降低10%。
