关键词： 神经网络模型   攻击方法   演化算法   人工智能技术   机器学习方法   黑箱   自然语言处理   学习模型  

摘要： 一、相关介绍目前,人工智能技术发展火热,其中离不开神经网络的快速崛起。作为目前在计算机视觉、强化学习、自然语言处理、语音识别等领域效果最佳的机器学习模型,神经网络以其强大的特征提取能力、数据拟合能力而著称。相比传统机器学习方法,神经网络不需要手工选择特征,其自组织、自学习的特点,更能够帮助模型捕捉到各类高维数据中隐含着的潜在规律。并且,越来越多的智能应用开始逐渐落地,例如智能驾驶系统中的视觉感知、人脸识别应用、智慧城市等等。值得注意的是,最近越来越多的研究证明,神经网络模型并非是一个足够鲁棒、安全的模型。

关键词： Human mobility patterns   Mobile phones   Machine learning   Data mining  

摘要： Human mobility patterns reflect many aspects of life, from the global spread of infectious diseases to urban planning and daily commute patterns. In recent years, the prevalence of positioning methods and technologies, such as the global positioning system, cellular radio tower geo-positioning, and WiFi positioning systems, has driven efforts to collect human mobility data and to mine patterns of interest within these data in order to promote the development of location-based services and applications. The efforts to mine significant patterns within large-scale, high-dimensional mobility data have solicited use of advanced analysis techniques, usually based on machine learning methods, and therefore, in this paper, we survey and assess different approaches and models that analyze and learn human mobility patterns using mainly machine learning methods. We categorize these approaches and models in a taxonomy based on their positioning characteristics, the scale of analysis, the properties of the modeling approach, and the class of applications they can serve. We find that these applications can be categorized into three classes: user modeling, place modeling, and trajectory modeling, each class with its characteristics. Finally, we analyze the short-term trends and future challenges of human mobility analysis.

关键词： Radiomics   Isocitrate dehydrogenase   Diffuse glioma   Magnetic resonance imaging   Machine learning  

摘要： Purpose Reliable and accurate predictive models are necessary to drive the success of radiomics. Our aim was to identify the optimal radiomics-based machine learning method for isocitrate dehydrogenase (IDH) genotype prediction in diffuse gliomas. Methods Eight classical machine learning methods were evaluated in terms of their stability and performance for pre-operative IDH genotype prediction. A total of 126 patients were enrolled for analysis. Overall, 704 radiomic features extracted from the pre-operative MRI images were analyzed. The patients were randomly assigned to either the training set or the validation set at a ratio of 2:1. Feature selection and classification model training were done using the training set, whereas the predictive performance and stability of the model were independently assessed using the validation set. Results Random Forest (RF) showed high predictive performance (accuracy 0.885 +/- 0.041, AUC 0.931 +/- 0.036), whereas neural network (NN) (accuracy 0.829 +/- 0.064, AUC 0.878 +/- 0.052) and flexible discriminant analysis (FDA) (accuracy 0.851 +/- 0.049, AUC 0.875 +/- 0.057) displayed low predictive performance. With regard to stability, RF also showed high robustness against data perturbation (relative standard deviations, RSD 3.87%). Conclusions RF is a promising machine learning method in predicting IDH genotype. Development of an accurate and reliable model can assist in the initial diagnostic evaluation and treatment planning for diffuse glioma patients.

关键词： CyberManufacturing systems   Security   Additive manufacturing   Machine learning  

摘要： CyberManufacturing system (CMS) is a vision for future manufacturing systems. The concept delineates a vision of advanced manufacturing system integrated with technologies such as Internet of Things, Cloud Computing, Sensors Network and Machine Learning. As a result, cyber-attacks such as Stuxnet attack will increase along with growing simultaneous connectivity. Now, cyber-physical attacks are new and unique risks to CMSs and modern cyber security countermeasure is not enough. To learn this new vulnerability, the cyber-physical attacks is defined via a taxonomy under the vision of CMS. Machine learning on physical data is studied for detecting cyber-physical attacks. Two examples were developed with simulation and experiments: 3D printing malicious attack and CNC milling machine malicious attack. By implementing machine learning methods in physical data, the anomaly detection algorithm reached 96.1% accuracy in detecting cyber-physical attacks in 3D printing process;random forest algorithm reached on average 91.1% accuracy in detecting cyber-physical attacks in CNC milling process.

关键词： process equipment   detection of abnormalities   predictive analytics   committee of classifiers   logistic regression  

摘要： -The article discusses a method for early detection and prediction of abnormality in operation of power-unit process equipment taking as an example the PTN 1100-350-17-4 turbine driven feedwater pump of a 300 MW power unit. The importance of the problem of predicting possible process equipment malfunctions at an early state of their occurrence is determined, and the specific features of solving it in the power industry are explained. The range of process equipment defects that can be efficiently detected using the predictive analytics methods is outlined. The fundamental assertion stating that the scope of analog and discrete measurements available in the process control system's set of computerized automation tools is sufficient for applying the predictive analytics methods is emphasized. Modern predictive analytics methods are briefly reviewed, and the specific features of model training algorithms are mentioned. Separate attention is paid to the problems of preparing initial data for training the model. The mathematical problem of modeling an abnormality indicator taking the values from 0 (normal operation) to 1 (abnormal operation) is formulated. In turn, this problem is formulated as the binary classification problem of attribute vectors characterizing the equipment state at the given moment of time. An original approach is suggested, which combines the multivariate state estimation technique (MSET), in which the degree of abnormality in a technical state is determined from the extent to which the Hotelling criterion exceeds a threshold level (which is automatically calculated in the algorithm), and machine learning methods, the use of which makes it possible to overcome a number of difficulties inherent in the MSET. For solving the problem of determining the composition of the most informative attributes from the values of which early development of an emergency can be detected, it is proposed to use an ensemble of regression models. A method for selecting the

关键词： Fault detection and location   Modular multilevel converter   Mixed kernel support tensor machine   Machine learning  

摘要： In this paper, the fault diagnosis and location (FDL) problem of the open circuit fault for modular multilevel converter (MMC) is investigated. A mixed kernel support tensor machine (MKSTM) is provided, and it's employed to improve the support tensor machine which is an important algorithm of machine learning. By extracting the characteristic data of ac current and internal circulation current in either normal operation or open-circuit fault, then training and classifying the obtained samples with MKSTM, FDL of MMC can be realized with the supplied algorithm. Finally, experimental results show that the classification accuracy of MKSTM algorithm is improved observably than single kernel function STM such as linear, Radial Basis function (RBF), sigmoid and polynomial types. Synchronously, the open-circuit fault can be effectively diagnosed and located with the proposed method. (C) 2018 Elsevier B.V. All rights reserved.

关键词： MACHINE learning   METHYLCYTOSINE   EPIGENETICS   DNA repair   ELECTRON-ion collisions  

摘要： Motivation: N-4-methylcytosine (4mC), an important epigenetic modification formed by the action of specific methyltransferases, plays an essential role in DNA repair, expression and replication. The accurate identification of 4mC sites aids in-depth research to biological functions and mechanisms. Because, experimental identification of 4mC sites is time-consuming and costly, especially given the rapid accumulation of gene sequences. Supplementation with efficient computational methods is urgently needed. Results: In this study, we developed a new tool, 4mCPred, for predicting 4mC sites in Caenorhabditis elegans, Drosophila melanogaster, Arabidopsis thaliana, Escherichia coli, Geoalkalibacter subterraneus and Geobacter pickeringii. 4mCPred consists of two independent models, 4mCPred_I and 4mCPred_II, for each species. The predictive results of independent and cross-species tests demonstrated that the performance of 4mCPred_I is a useful tool. To identify position-specific trinucleotide propensity (PSTNP) and electron-ion interaction potential features, we used the F-score method to construct predictive models and to compare their PSTNP features. Compared with other existing predictors, 4mCPred achieved much higher accuracies in rigorous jackknife and independent tests. We also analyzed the importance of different features in detail.

关键词： 平滑肌肉瘤   表达谱   机器学习   分子分型   预测标签  

摘要： [目的]应用机器学习方法,利用TCGA与GEO数据,构建平滑肌肉瘤分子分型预测标签。[方法]收集TCGA和GSE45510数据,建立机器学习模型,利用pamr包预测标签,将预测亚型与实际值对比以评估预测效果。[结果]经测试,阈值取6.639时,分类准确率为87.8%,相关基因159个。[结论]利用机器学习方法构建平滑肌肉瘤分子分型预测标签,可以用于预测平滑肌肉瘤分子亚型,提高诊断精度,指导未来平滑肌肉瘤亚型特异性靶向治疗,推动精准治疗发展。

关键词： 土地利用/土地覆被分类   Landsat   OLI影像   机器学习   人工神经网络   决策树   支持向量机   随机森林   湟水流域  

摘要： 针对高海拔复杂地形区地貌类型复杂、多样,沟壑纵横、地形破碎等特点,研究快速、有效的土地利用/土地覆被分类方法对土地利用信息获取及更新是非常重要的。以位于黄土高原向青藏高原过渡带的湟水流域为研究区域,基于Landsat 8 OLI影像数据、DEM数据,并结合各种专题特征,在对研究区进行地理分区的基础上,采用人工神经网络、决策树、支持向量机和随机森林4种机器学习方法进行土地利用信息提取并进行精度评价,探索适合于复杂地形区最优的分类方法。研究结果表明:随机森林和决策树的分类精度明显高于支持向量机和人工神经网络。其中随机森林方法的分类精度最高,总体分类精度达85.65%,Kappa系数达0.84。在上述分类基础上,选择随机森林分类方法对Landsat 8全色与多光谱影像融合数据进行进一步的分类研究,总体分类精度达到86.49%,Kappa系数达0.85。这表明随机森林分类方法在保证分类精度的同时又能获得较高的分类效率,对于复杂地形区土地利用信息提取是非常有效的,数据融合在一定程度上提高了分类精度。