关键词： Cloud   Virtual machine   Volatile memory   Malware   Linux   Detection   Machine learning   Feature extraction   Volatility  

摘要： Most organizations today use cloud-computing environments and virtualization technology. Linux-based clouds are the most popular cloud environments among organizations, and thus have become the target of cyber-attacks launched by sophisticated malware. Existing malware detection solutions for Linux-based VMs are installed and operated on the VM itself and are considered untrusted since malware can detect, interfere with, and even evade them. Thus, Linux cloud-based environments remain exposed to various malware-based attacks. This paper presents the first trusted framework for detecting unknown malware in Linux VM cloud-environments. Our framework acquires volatile memory dumps from the inspected VM by querying the hypervisor in a trusted manner and overcoming malware's ability to detect the security mechanism and evade detection. Then, using machine-learning algorithms we leverage informative traces (our 171 proposed features) from different parts of the VM's volatile memory. The framework was evaluated in seven rigorous experiments, on a total of 21,800 volatile memory dumps taken from two widely used virtual servers (10,900 from each server) during the execution of a diverse yet representative collection of benign and malicious Linux applications. Notably, the results show that our proposed framework can accurately (with high TPRs and low FPRs): (a) detect unknown malware (b) detect new unknown malware from unseen malware categories, which is a critical ability for coping with new malware trends and phenomena;(c) categorize an unknown malware by its attack category;(d) detect unknown malware on an unknown virtual-server;and lastly (e) detect fileless malware, a critical capability demonstrating the ability to detect substantially different attack modus operandi. (C) 2021 Elsevier B.V. All rights reserved.

关键词： Short-term load forecasting   Workdays and weekends   Fuzzy time series   Multi-objective optimization  

摘要： Countries around the globe have introduced renewable energies (RE) and minimized the dependency of fossil resources in power systems to address extensive environmental risks. However, such large-scale energy transitions pose a great challenge to power systems due to the volatility of RE. Meanwhile, power demand is increasing over time and it shows temporal characteristics, such as seasonal and peak-valley patterns. Whether the future power system with a larger proportion of RE can meet the surging but fluctuated electricity demand remains problematic. Previous studies on short-term load forecasting focused more on forecasting accuracy than stability. Further, there is a relative paucity of research into temporal patterns. In order to fill in these research gaps, this paper proposes a fuzzy theory-based machine learning model for workdays and weekends short-term load forecasting. Fuzzy time series (FTS) is applied for data mining and back propagation (BP) neural network is used as the main predictor for short-term load forecasting. To exploit the trade-offs between forecasting stability and accuracy, multi-objective optimization is applied to modify the parameters of BP. Moreover, an interval forecasting architecture with several statistical tests is constructed to address forecasting uncertainties. Short-term load data from Victoria in Australia is selected as a case study. Results demonstrate that the proposed method can significantly boost forecasting stability and accuracy, and help strategy making in the field of energy and electricity system management and planning. (c) 2021 The Author. Published by Elsevier B.V. This is an open access article under the CC BY license (http:// ***/licenses/by/4.0/).

关键词： Strain   Temperature sensors   Fiber gratings   Temperature measurement   Sensors   Boosting   Training   Bragg grating   ensemble methods   machine learning   optical fiber sensor   strain   temperature measurement  

摘要： The biggest challenge of using fiber Bragg grating (FBG) based sensors is the cross-sensitivity between the strain and temperature effects on FBG. In this letter, we demonstrate the ability of machine learning (ML) methods to discriminate between the strain and temperature effects on FBG sensors on a single measurement of change in the Bragg wavelength. Spectral data are collected using an FBG interrogation system at various strain and temperature conditions and are applied to different ML methods to determine the strain and temperature effects. We further simulate FBG with the same strain and temperature conditions using VPIphotonics. For comparison, the same ML methods are applied to both simulated and experimentally collected data. The experimental results reveal that our proposed model can predict strain and temperature with 90% accuracy on a single measurement of Bragg wavelength. We also demonstrate the stability of the model by comparing the testing and training errors of the applied ML methods. Therefore, our proposed technique reduces the cost and complexity associated with the existing FBG-based sensor system.

关键词： Terahertz spectroscopy   Bisphenol   Quantitative analysis   Analogue mixtures   SVR  

摘要： Quantitative analysis of complex mixtures is a great challenge for spectral analysis. Bisphenol A (BpA) is a chemical predominantly used in manufacturing and is being replaced by other analogs due to its potential toxicity. Reliability methods is hence crucial for identification and quantification of bisphenol mixtures. In this study we present an attractive strategy for composition determination of BpA incorporated in its analogue mixtures. Terahertz spectra of four bisphenol components are analyzed using machine learning method (SVR) to learn the underlying model of the frequency against the target concentration of BpA in mixtures. The learned mode predicts the concentrations of the unknown samples with decision coefficient R2 = 0.98. Absorption spectra for bisphenols mixtures were successfully reconstructed by a hold-out validation scheme. The results indicate the terahertz spectroscopy in combination with SVR is robust and accurate in mixture quantitative analysis and should play a significant role for industrial applications in the future.

关键词： 医学图像   图像压缩   机器学习  

摘要： 在医学研究中会产生大量需要储存、传输的图像数据,如计算机断层扫描(CT)等.医院要处理海量患者的所有数据相当困难,必须对这些数据进行压缩.随着人工智能技术的发展,应用机器学习的医学图像数据压缩方法成为新的研究热点.文中首先总结了传统医学图像数据压缩方法;然后基于机器学习的医学图像数据压缩方法,对各类方法针对脑部CT、肺部CT等不同医学图像的压缩效果进行对比,并对这些方法在压缩率、算法复杂度、重构质量等方面的优缺点进行系统的总结,指出机器学习和感兴趣区域结合的方法可以在享受有损压缩带来的高压缩率的同时保留重要区域的特征信息,比其他压缩方法更贴合医学数据的压缩要求;最后展望了该领域的未来方向和挑战.

关键词： KNN   机器学习   青藏高原   多年冻土   土壤水热过程  

摘要： 利用唐古拉站2004—2012年气象观测资料,基于KNN算法,结合机器学习思想,建立了一个气象回归模型,模拟了2005年唐古拉地区表层土壤水热变化趋势,结合实测数据,将模拟值与观测值进行对比,并对模型模拟效果进行了评估。结果表明:KNN模型能够较好地模拟活动层表层土壤水热状况,各层土壤温度的模拟值与观测值的相关系数均在0.99以上,均方根误差在1.25℃以内;不同深度土壤水分的模拟值与观测值的相关系数均在0.95以上,均方根误差在0.02 m^(3)·m^(-3)以内。总体上,KNN模型能够对青藏高原多年冻土区唐古拉地区表层土壤水热状况进行较为精确地模拟,该模型对于青藏高原其他地区的适用性有待进一步研究验证。

关键词： Machine learning   Concrete creep   Support vector regression   Sensitivity   Maximum information correlation  

摘要： Although machine learning algorithms to predict the mechanical properties of concrete have been studied extensively, most of the research focused on the prediction of the strength of concrete and only a few studies have focused on concrete creep. This paper analyzed the maximum information correlation (MIC) between concrete creep influence parameters based on the updated Infrastructure Technology Institute of Northwestern University (NU-ITI) database, and the parameters in the database were adopted in the classical creep prediction models for calculation. Three machine learning algorithms (MLAs)-back-propagation artificial neural network (BPANN), support vector regression (SVR), and extreme learning machine (ELM)-were trained with the NU-ITI database to model concrete creep. The SVR-based model achieved high predictive accuracy. Sensitivity analysis of the parameters and feature selection of concrete creep were carried out based on the SVR and the Sobol method. By retraining the SVR model after feature selection, it was demonstrated that low-sensitivity and strongly correlated parameters will increase the robustness of machine learning models.

关键词： 机器学习方法   信用风险   评估  

摘要： 基于信用风险评估的文献随着时间的推移以及技术创新的不断更新,相关机器学习方法运用近年来也与时俱进。基于此,主要从传统的机器学习方法、近现代机器学习方法以及近三年来最新研究三个方面进行信用风险评估领域文献梳理与评述,归纳现有机器学习方法存在的问题和不足,并对未来机器学习方法在金融风控领域研究方向做出展望,提出发展趋势。

关键词： 图卷积神经网络   深度学习   高维组学   分类预测  

摘要： 目的探讨图卷积神经网络(graph convolutional neural network,GCNN)利用PPI网络对组学数据的分类预测效能。方法通过模拟实验和实例研究,对GCNN、随机森林、支持向量机和多层感知机共四种方法的分类效果进行比较。结果模拟实验结果显示,即便在样本量和网络中节点数量较少时,GCNN的分类效能也明显优于其他三种方法,并且随着节点数量的增加而不断提高。实例研究表明,利用STRING网络,GCNN的分类效能最优。结论GCNN在组学数据的研究中极具潜力,值得进一步研究。

关键词： Interpretable machine learning   SHAP   Tree-based model   Species distribution model   Macroinvertebrate   EPT taxa  

摘要： Species distribution models (SDMs), in which species occurrences are related to a suite of environmental variables, have been used as a decision-making tool in ecosystem management. Complex machine learning (ML) algorithms that lack interpretability may hinder the use of SDMs for ecological explanations, possibly limiting the role of SDMs as a decision-support tool. To meet the growing demand of explainable MLs, several interpretable ML methods have recently been proposed. Among these methods, SHaply Additive exPlanation (SHAP) has drawn attention for its robust theoretical justification and analytical gains. In this study, the utility of SHAP was demonstrated by the application of SDMs of four benthic macroinvertebrate species. In addition to species responses, the dataset contained 22 environmental variables monitored at 436 sites across five major rivers of South Korea. A range of ML algorithms was employed for model development. Each ML model was trained and optimized using 10-fold cross-validation. Model evaluation based on the test dataset indicated strong model performance, with an accuracy of >= 0.7 in all evaluation metrics for all MLs and species. However, only the random forest algorithm showed a behavior consistent with the known ecology of the investigated species. SHAP presents an integrated framework in which local interpretations that incorporate local interaction effects are combined to represent the global model structure. Consequently, this framework offered a novel opportunity to assess the importance of variables in predicting species occurrence, not only across sites, but also for individual sites. Furthermore, removing interaction effects from variable importance values (SHAP values) clearly revealed non-linear species responses to variations in environmental variables, indicating the existence of ecological thresholds. This study provides guidelines for the use of a new interpretable method supporting ecosystem management.