无线网络的分布式入侵检测模型研究

对无线网络安全方面的问题和无线网络的特点进行了概述,分析了无线网络的入侵检测模型进行了分析,提出了一个分分布式的无线网络入侵检测系统模型,并给出了具体的拓扑结构设计,通过实验,证明这个入侵检测系统的设计和实现具有很好的安全性,其安全保障具有一定的实用性,值得推广。     

基于地面先验的3D目标检测算法

为了提高自动驾驶汽车环境感知的性能,增强单目相机对障碍物三维和边界信息的感知能力,提出了一种基于地面先验的3D目标检测算法。基于优化的中心网络(CenterNet)模型,以DLA(deep layer aggregation)为主干网络,增加目标3D边沿框中心点冗余信息预测。根据自动驾驶场景的地面先验信息,结合针孔相机模型,获取目标3D中心深度信息,以优化深度网络学习效果。使用KITTI 3D数据集评测算法性能,结果表明:在保证2D目标检测准确性的基础上,该算法运行帧率约20 fps,满足自动驾驶感知实时性要求;同时相比于CenterNet模型,在平均方位角得分(average orientation score)和鸟视图平均准确率(bird eye view AP)上分别有4.4和4.4%的性能提升。因而,该算法可以提高自动驾驶汽车对障碍物三维和边界信息的感知能力。     

基于可编程协议无关报文处理的分布式拒绝服务攻击检测

传统软件定义网络（SDN）中的分布式拒绝服务（DDoS）攻击检测方法需要控制平面与数据平面进行频繁通信,这会导致显著的开销和延迟,而目前可编程数据平面由于语法无法实现复杂检测算法,难以保证较高检测效率。针对上述问题,提出了一种基于可编程协议无关报文处理（P4）可编程数据平面的DDoS攻击检测方法。首先,利用基于P4改进的信息熵进行初检,判断是否有可疑流量发生;然后再利用P4提取特征只需微秒级时长的优势,提取可疑流量的六元组特征导入数据标准化—深度神经网络（data standardization-deep neural network,DS-DNN）复检模块,判断其是否为DDoS攻击流量;最后,模拟真实环境对该方法的各项评估指标进行测试。实验结果表明,该方法能够较好地检测SDN环境下的DDoS攻击,在保证较高检测率与准确率的同时,有效降低了误报率,并将检测时长缩短至毫秒级别。     

局部特征表征的6D位姿估计算法

为解决有纹理模型在遮挡条件下6D位姿估计精确度不高的问题,提出了一种局部特征表征的端到端6D位姿估计算法。首先为了得到准确的定位信息,提出了一个空间—坐标注意力机制（spatial and coordinate attention）,通过在YOLOv5网络中加入空间—坐标注意力机制和加权双向特征金字塔网络（bidirectional feature pyramid network）,YOLOv5-CBE算法的精确度（precision）、召回率（recall）、平均精度均值（mAP@0.5）分别提升了3.6%、2.8%、2.5%,局部特征中心点坐标误差最高提升了25%;然后用 YOLOv5-CBE算法检测局部特征关键点,结合3D Harris关键点通过奇异值分解法（singular value decomposition）计算模型的6D位姿,最高遮挡70%的情况下仍然可以保证二维重投影精度（2D reprojection accuracy）和ADD度量精度（ADD accuracy）在95%以上,具有较强的鲁棒性。     

基于群智感知的街景变化检测方法

移动群智感知数据包含的图像和时空情境信息可用于检测街景图像变化,但是群智感知数据通常是低质和不规范的。为了准确检测街景发生的变化,主要解决由拍摄视角差异引起的数据低质问题。首先,针对大视差问题采用图像配准方法初步对齐图像并提取出配准特征点;然后,基于配准特征点分布从图像中提取感兴趣区域;随后,针对差值图像的误检内容,提出基于面积和多特征点的筛选法去除误检区域;最后,结合边缘检测和超像素分割算法提取完整的变化对象。与MDFNet方法进行比较,实验结果显示：当街景发生变化时,该方法的F1-measure值为55.8%,增长6%,错误率为10.8%,降低24%;当街景无变化时,该方法的错误率为2.8%,下降28%。     

基于强化学习的无线传感器网络入侵检测攻防博弈研究

无线传感器网络易遭到各种内部攻击,入侵检测系统需要消耗大量能量进行攻击检测以保障网络安全。针对无线传感器网络入侵检测问题,建立恶意节点（malicious node,MN）与簇头节点（cluster head node,CHN）的攻防博弈模型,并提出一种基于强化学习的簇头入侵检测算法——带有近似策略预测的策略加权学习算法（weighted policy learner with approximate policy prediction,WPL-APP）。实验表明,簇头节点采用该算法对恶意节点进行动态检测防御,使得博弈双方快速达到演化均衡,避免了网络出现大量检测能量消耗和网络安全性能的波动。     

基于矩阵分解双线性池化的多模态融合虚假新闻检测

现有的大多数虚假新闻检测方法将视觉和文本特征串联拼接,导致模态信息冗余并且忽略了不同模态信息之间的相关性。为了解决上述问题,提出一种基于矩阵分解双线性池化的多模态融合虚假新闻检测算法。首先,该算法将多模态特征提取器捕捉的文本和视觉特征利用矩阵分解双线性池化方法进行有效融合,然后与虚假新闻检测器合作鉴别虚假新闻;此外,在训练阶段加入了事件分类器来预测事件标签并去除事件相关的依赖。在Twitter和微博两个多模态谣言数据集上进行了对比实验,证明了该算法的有效性。实验结果表明提出的模型能够有效地融合多模态数据,缩小模态间的异质性差异,从而提高虚假新闻检测的准确性。     

基于ADASYN-SFS-RF的欺诈检测模型泛化性能提升及可解释性研究

针对行业欺诈行为形式多样、操作隐蔽,且数据分布极端不平衡等问题,研究采用ADASYN（adaptive synthetic sampling approach for imbalanced learning）算法将分类决策边界向困难的实例进行自适应移动实现数据扩增,以解决不平衡数据造成的过拟合问题。采用基于随机森林的序列向前搜索策略算法筛选出最优特征子集对欺诈进行检测,降低ADASYN算法添加噪声数据对分类边界确定的影响,构建欺诈检测模型,并使用LIME对模型检测结果作出局部解释,提高模型的使用价值。实验表明,该模型可以较好地克服传统欺诈检测模型对多数类样本误分类的缺陷,有助于提高行业对交易欺诈行为识别的效率。同时,通过LIME对模型检测出的随机样本进行有效解析,便于决策者对算法模型的检测结果作出实证分析,起到明显的预警及决策参考价值。     

Using CANARY event detection software for water quality analysis in the Milwaukee River

Urban water sources are susceptible to various contamination events as a result of natural, accidental, and human-induced occurrences. An early warning monitoring system provides timely information on changes in urban water quality. In this study, an analysis was made with CANARY event detection software (EDS) to monitor water quality parameters in river water and to identify the onset of anomalous water quality periods. Water quality signals including pH, conductivity, and turbidity from the Milwaukee River over specified periods during the summer season of 2018–2020 were employed as inputs to event detection algorithms in CANARY. The data analysis results show that CANARY can be useful as an early warning system for monitoring contamination in urban water sources and help to identify abnormal conditions quickly. The sensibility of the model relies on optimizing the configuration parameters, which involves selecting the ideal set of parameters for the event detection algorithm and adjusting the BED parameters to increase or decrease the probability of generating an alarm. The number of events reported between the Linear Prediction Correction Filter (LPCF) and Multivariate Nearest Neighbor (MVNN) algorithms varied as a result of different residual calculation mechanisms. Climate factors that contributed to the abnormal water quality events in the river were examined. The analysis of rainfall on water quality was carried out using a statistical method by determining whether there is a significant difference (p-value) between the seasonal mean water quality data and the mean value of water parameters during the sampling duration. Regression analysis was also performed to estimate the best model that describes the relationship between each of the water quality parameters and temperature.     

Automatic Channel Detection Using DNN on 2D Seismic Data

Geologists interpret seismic data to understand subsurface properties and subsequently to locate underground hydrocarbon resources. Channels are among the most important geological features interpreters analyze to locate petroleum reservoirs. However, manual channel picking is both time consuming and tedious. Moreover, similar to any other process dependent on human intervention, manual channel picking is error prone and inconsistent. To address these issues, automatic channel detection is both necessary and important for efficient and accurate seismic interpretation. Modern systems make use of real-time image processing techniques for different tasks. Automatic channel detection is a combination of different mathematical methods in digital image processing that can identify streaks within the images called channels that are important to the oil companies. In this paper, we propose an innovative automatic channel detection algorithm based on machine learning techniques. The new algorithm can identify channels in seismic data/images fully automatically and tremendously increases the efficiency and accuracy of the interpretation process. The algorithm uses deep neural network to train the classifier with both the channel and non-channel patches. We provide a field data example to demonstrate the performance of the new algorithm. The training phase gave a maximum accuracy of 84.6% for the classifier and it performed even better in the testing phase, giving a maximum accuracy of 90%.