传统物联网设备指纹提取方法通常将流量中的隐私数据用于生成设备指纹并且采用手工设计特征的方式,在形成安全隐患的同时限制了模型的性能。针对上述问题,提出一种基于设备行为的物联网设备指纹深度提取方法(IoT device deep fingerprin...传统物联网设备指纹提取方法通常将流量中的隐私数据用于生成设备指纹并且采用手工设计特征的方式,在形成安全隐患的同时限制了模型的性能。针对上述问题,提出一种基于设备行为的物联网设备指纹深度提取方法(IoT device deep fingerprint extraction,IDFE)。IDFE将网络流量pcap文件划分为多个会话(sessions),并提取非隐私信息构建会话信息矩阵,设计了会话信息矩阵不同信息序列之间的依赖关系和会话数据包之间的时序依赖关系建模方法和融合方法,利用设计的全卷积Transformer提取融合后的会话特征矩阵中设备行为特征并生成设备指纹。在UNSW和YourThings两个公开数据集上进行了广泛的实验,验证了该方法的有效性。展开更多
Within a honeybee population, due to polyandry, there are super-sister and half-sister relations, thus many sub-families exist. For the Chinese honeybee (Apis cerana cerana F.) a phylogenetic dendrogram has been const...Within a honeybee population, due to polyandry, there are super-sister and half-sister relations, thus many sub-families exist. For the Chinese honeybee (Apis cerana cerana F.) a phylogenetic dendrogram has been constructed in which 4 sub-families are clustered based on DNA fingerprint patterns. And it has been observed that each kind of workers is distributed to several different sub-families.展开更多
文摘传统物联网设备指纹提取方法通常将流量中的隐私数据用于生成设备指纹并且采用手工设计特征的方式,在形成安全隐患的同时限制了模型的性能。针对上述问题,提出一种基于设备行为的物联网设备指纹深度提取方法(IoT device deep fingerprint extraction,IDFE)。IDFE将网络流量pcap文件划分为多个会话(sessions),并提取非隐私信息构建会话信息矩阵,设计了会话信息矩阵不同信息序列之间的依赖关系和会话数据包之间的时序依赖关系建模方法和融合方法,利用设计的全卷积Transformer提取融合后的会话特征矩阵中设备行为特征并生成设备指纹。在UNSW和YourThings两个公开数据集上进行了广泛的实验,验证了该方法的有效性。
文摘Within a honeybee population, due to polyandry, there are super-sister and half-sister relations, thus many sub-families exist. For the Chinese honeybee (Apis cerana cerana F.) a phylogenetic dendrogram has been constructed in which 4 sub-families are clustered based on DNA fingerprint patterns. And it has been observed that each kind of workers is distributed to several different sub-families.