ID 原文 译文
18885 其次，为了克服优化模型中难以准确掌握系统状态转移概率及状态空间过大的问题，该文提出了一种基于强化学习框架的VNF智能迁移学习算法， Secondly, in order to overcome the issue of having difficulties in acquiring the accurate transition probabilities of the system states and the excessive state space inthe optimization model, a VNF intelligent migration learning algorithm based on reinforcement learning framework is proposed.
18886 该算法通过卷积神经网络(CNN)来近似行为值函数，从而在每个离散的时隙内根据当前系统状态为每个网络切片制定合适的VNF迁移策略及CPU资源分配方案。 The algorithm approximates the behavior value function by Convolutional NeuralNetwork (CNN), so as to formulate a suitable VNF migration strategy and CPU resource allocation scheme foreach network slicing according to the current system state in each discrete time slot.
18887 仿真结果表明，所提算法在有效地满足各切片QoS需求的同时，降低了基础设施的平均能耗。 The simulation results show that the proposed algorithm can effectively meet the QoS requirements of each slice while reducing the average energy consumption of the infrastructure.
18888 针对网络流量异常检测过程中提取的流量特征准确性低、鲁棒性差导致流量攻击检测率低、误报率高等问题，该文结合堆叠降噪自编码器(SDA)和softmax，提出一种基于深度特征学习的网络流量异常检测方法。 In view of the problems of low attack detection rate and high false positive rate caused by pooraccuracy and robustness of the extracted traffic features in network traffic anomaly detection, a network trafficanomaly detection method based on deep features learning is proposed, which is combined with StackedDenoising Autoencoders (SDA) and softmax.
18889 首先基于粒子群优化算法设计SDA结构两阶段寻优算法：根据流量检测准确率依次对隐藏层层数及每层节点数进行寻优，确定搜索空间中的最优SDA结构，从而提高SDA提取特征的准确性。 Firstly, a two-stage optimization algorithm is designed based onparticle swarm optimization algorithm to optimize the structure of SDA, the number of hidden layers andnodes in each layer is optimized successively based on the traffic detection accuracy, and the optimal structureof SDA in the search space is determined, improving the accuracy of traffic features extracted by SDA.
18890 然后采用小批量梯度下降算法对优化的SDA进行训练，通过最小化含噪数据重构向量与原始输入向量间的差异，提取具有较强鲁棒性的流量特征。 Secondly, the optimized SDA is trained by the mini-batch gradient descent algorithm, and the traffic features with strong robustness are extracted by minimizing the difference between the reconstruction vector of the corrupted data and the original input vector.
18891 最后基于提取的流量特征对softmax进行训练构建异常检测分类器，从而实现对流量攻击的高性能检测。 Finally, softmax is trained by the extracted traffic features toconstruct an anomaly detection classifier for detecting traffic attacks with high performance.
18892 实验结果表明：该文所提方法可根据实验数据及其分类任务动态调整SDA结构，提取的流量特征具有更高的准确性和鲁棒性，流量攻击检测率高、误报率低。 The experimental results show that the proposed method can adjust the structure of SDA based on the experimental data and its classification tasks, extract traffic features with a higher accuracy and robustness, and detect traffic attacks with high detection rate and low false positive rate.
18893 正交频分复用(OFDM)已被广泛应用于无线通信系统，其数据传输安全具有一定的实际意义。 Orthogonal Frequency Division Multiplexing(OFDM) is widely used in wireless communication systems, and its data transmission security has certain practical significance.
18894 该文提出了一种双重加密方案，采用神经网络生成置乱矩阵实现第1次加密，通过基于Logistic映射与Sine映射的复合离散混沌系统产生的混沌序列进行第2次加密。该双重加密方案极大提升了OFDM通信系统的保密性，可以有效地防止暴力攻击。 A double encryption scheme is proposed which enhances the confidentiality of the OFDM communication system and can prevent brute force attacks significantly. Specifically, the first encryption is achieved by using neural network to generate the scrambling matrix, and the second encryption is implemented by chaotic sequence generating by composite discrete chaotic system based on Logistic mapping and Sine mapping.