ID 原文 译文
3003 最后，利用多个真实复杂网络数据集进行数值算例。 Finally, several real complex network datasets were used for numerical examples calculation.
3004 实验结果表明，与其他基线方法相比，所提方法具有更优的精度与效率。 Experimental results show that the proposed algorithm is more accurate and efficient than other baseline methods.
3005 针对软件定义网络（SDN）现有转发验证机制因嵌入额外的分组字段所带来的通信开销大的问题，提出基于端址重载的包转发验证机制。 Aiming at the problem that the existing forwarding verification mechanisms in software-defined networking(SDN) incur significant communication overhead caused by embedding additional packet fields, a packet forwarding ver-ification mechanism based on port address overloading was proposed,
3006 其核心思想是入口交换机重构数据分组端口和地址信息实现端址重载，下游交换机基于重载的端址信息实现数据分组的概率验证，控制器统计路径中节点验证有效和无效的数据分组信息并定位异常； which key idea was the ingress switch implement-ed port address overloading by reconstructing port and address of packet, downstream switches executed packet proba-bilistic verification based on overloading port address, and the controller acquired valid and invalid packet statistics of node verification in the path and localized anomaly.
3007 理论分析给出了恶意注入与丢弃攻击异常检测阈值； Anomaly detection threshold of malicious injecting and dropping packets was presented by theoretical analysis.
3008 最后实现了所提机制并对其进行了评估。 Finally, the proposed scheme was implemented and evaluated.
3009 实验结果表明，所提机制以引入不超过 10%的转发时延、低于 8%的吞吐率损失实现高效转发及有效的异常定位。 Experiments demonstrate the proposed scheme achieves efficient forwarding and effective anomaly localization with less than 10% ofadditional forwarding delays and less than 8% of throughput degradation.
3010 为识别多入多出正交空时分组码（MIMO-OSTBC）系统所采用的调制样式，提出了一种基于一维卷积神经网络（1D-CNN）的协作调制识别算法。 To recognize the modulation style adopted in multiple-input-multiple-output orthogonal space-time block code(MIMO-OSTBC) systems, a cooperative modulation recognition algorithm based on the one-dimensional convolutionalneural network (1D-CNN) was proposed.
3011 首先，采用迫零盲均衡来提升不同调制信号间区分度，并选用天然无损的同相正交（I/Q）信号作为浅层特征； With the lossless I/Q signal selected as shallow features, the zero-forcing blindequalization was first leveraged to improve the discrimination of different modulation signals.
3012 然后，设计并训练基于 1D-CNN 的识别模型，从浅层特征中提取深层特征； Then the 1D-CNN recog-nition model was devised and trained to extract deep features from shallow ones.