基于深度强化学习的恶意软件混淆对抗样本生成

设计一种PE格式恶意软件混淆对抗样本生成模型。利用深度强化学习算法,实现对恶意软件的自动混淆。通过加入历史帧和LSTM神经网络结构的方法使深度强化学习模型具有记忆性。对比实验表明,该恶意软件变种在基于机器学习的检测模型上的逃逸率高于现有研究,在由918个PE格式恶意软件组成的测试集上达到39.54%的逃逸率。     

一种识别病毒和蠕虫的算法

对现有的恶意软件检测算法进行研究之后发现,某些检测算法只能检测一种恶意软件,并且部分传统的检测算法在检测恶意程序时漏检率偏高；针对目前现有的检测算法缺乏综合性检测能力的短板,在此文中提出了一种新的检测算法,该检测算法具有一定的综合检测能力；新算法的思路如下:第一步区分某种软件是恶意软件还是非恶意软件,如果是恶意软件则提取其特征码,然后使用决策树根据恶意软件的特征码对恶意软件进行识别和分类,如果存在特征码不能识别的恶意软件,那么再根据病毒和蠕虫的特征使用相似性计算算法对未知的恶意软件进行相似性计算,最后使用决策系统对相似性算法计算的结果进行决策,该恶意软件是病毒还是蠕虫；将相似性计算算法,决策树和决策系统在检测恶意软件算法中进行应用是文章的创新之处。     

一种基于多特征集成学习的恶意代码静态检测框架

伴随着互联网的普及和5G通信技术的快速发展,网络空间所面临的威胁日益增大,尤其是恶意软件的数量呈指数型上升,其所属家族的变种爆发式增加.传统的基于人工签名的恶意软件的检测方式速度太慢,难以处理每天数百万计新增的恶意软件,而普通的机器学习分类器的误报率和漏检率又明显过高.同时恶意软件的加壳、混淆等对抗技术对该情况造成了更大的困扰.基于此,提出一种基于多特征集成学习的恶意软件静态检测框架.通过提取恶意软件的非PE(Portable Executable)结构特征、可见字符串与汇编码序列特征、PE结构特征以及函数调用关系5部分特征,构建与各部分特征相匹配的模型,采用Bagging集成和Stacking集成算法,提升模型的稳定性,降低过拟合的风险.然后采取权重策略投票算法对5部分集成模型的输出结果做进一步聚合.经过测试,多特征多模型聚合的检测准确率可达96.99%,该结果表明:与其他静态检测方法相比,该方法具有更好的恶意软件鉴别能力,对加壳、混淆等恶意软件同样具备较高的识别率.     

改进的基于DNN的恶意软件检测方法

当前基于深度学习的恶意软件检测技术由于模型结构及样本预处理方式不够合理等原因,大多存在泛化性较差的问题,即训练好的恶意软件检测模型对不属于训练样本集的恶意软件或新出现的恶意软件的检出效果较差。提出一种改进的基于深度神经网络（Deep Neural Network,DNN）的恶意软件检测方法,使用多个全连接层构建恶意软件检测模型,并引入定向Dropout正则化方法,在模型训练过程中对神经网络中的权重进行剪枝。在Virusshare和lynx-project样本集上的实验结果表明,与同样基于DNN的恶意软件检测模型DeepMalNet相比,改进方法对恶意PE样本集的平均预测概率提高0.048,对被加壳的正常PE样本集的平均预测概率降低0.64。改进后的方法具有更好的泛化能力,对模型训练样本集外的恶意软件的检测效果更好。     

基于带权欧拉距离的PE文件壳检测技术

越来越多的恶意软件出现在网络上。恶意软件作者通过网络将软件中的恶意代码植入用户的电脑中,从而达到诸如获得用户名与密码的非法目的。为了阻止它们对用户电脑的侵害,软件分析人员必须分析恶意软件的工作原理。但是,如果这些恶意软件加壳,那么分析它们就会变得非常困难,因此必须对他们进行脱壳。脱壳的第一步即检测这些恶意软件是否加壳。本文通过对未加壳和已经加壳的软件PE头部进行分析与比较,提出了带权欧拉距离PE文件壳检测(PDWED)算法,其中包括构造一个含有10个元素的向量,并为每个向量中每个元素分配一个权重值,计算向量的带权欧拉距离。实验结果表明,PDWED能够比较快速而又准确地检测软件是否加壳。     

基于Siamese架构的恶意软件隐藏函数识别方法北大核心CSCD

目前,隐藏技术已被普遍应用于恶意软件中,以避免反病毒引擎的检测及研究人员的反向分析,所以有效识别恶意软件中的隐藏函数对于恶意软件代码检测和深度分析具有重要意义。但在该领域上,现有方法不同程度都存在一些问题,如无法取得高准确性、对样本量少或者样本类别分布不平衡的数据集的鲁棒性较差等。为实现实用的针对恶意软件隐藏函数的检测方法,文章提出一种新颖的基于Siamese架构的识别方法来检测隐藏函数的类型。该方法可以有效提高隐藏函数识别的准确性,Siamese架构的引入改善了小样本量数据集鲁棒性差的问题。针对从恶意软件中提取的15种常见类型的隐藏函数的数据集进行实验,结果表明,该方法生成的嵌入向量较嵌入神经网络SAFE具有更好的质量,该方法较几种常用的隐藏函数检测工具有更高的检测精度。     

基于改进决策树分类的Android恶意软件检测

针对目前Android手机恶意软件越来越多的问题,本文在现有研究的基础上,设计了一个Android恶意软件检测框架。该框架通过提取Android应用程序的特征属性,结合Fisher Score、信息增益和卡方检验三种特征选择方法,对属性特征进行预处理,然后利用恶意检测模块中的改进决策树算法对软件进行分类。通过实验仿真,结果表明使用该检测框架检测恶意软件具有较低的误报率和较高的精确度。     

一种采用免疫原理的恶意软件检测方法

针对现有恶意软件检测方法的不足,提出一种采用免疫原理的恶意软件检测方法.该方法采用程序运行时产生的IRP请求序列作为抗原,定义系统中的正常程序为自体、恶意程序为非自体,通过选定数量的抗体,采用人工免疫原理识别非自体.实验结果表明,此方法在恶意软件的检测方面具有较高的准确率,且误报和漏报率较低.     

决策树方法在恶意DLL文件检测中的应用

本文对现有恶意DLL文件注入技术和PE结构进行了分析,提出了一种检测恶意DLL文件的新方法。通过分析DLL文件的文件属性和PE文件字段值,利用决策树方法中的C4.5算法构造恶意DLL文件检测模型,并且通过实验验证该检测模型的检测效率。     

一种恶意软件行为分析系统的设计与实现

基于虚拟化技术的恶意软件行为分析是近年来出现的分析恶意软件的方法。该方法利用虚拟化平台良好的隔离性和控制力对恶意软件运行时的行为进行分析,但存在两方面的不足：一方面,现有虚拟机监视器（Virtual Machine Monitor,VMM）的设计初衷是提高虚拟化系统的通用性和高效性,并没有充分考虑虚拟化系统的透明性,导致现有的VMM很容易被恶意软件的环境感知测试所发现。为此,提出一种基于硬件辅助虚拟化技术的恶意软件行为分析系统——THVA。THVA是一个利用了安全虚拟机（SVM）、二级页表（NPT）和虚拟机自省等多种虚拟化技术完成的、专门针对恶意软件行为分析的微型VMM。实验结果表明,THVA在行为监控和反恶意软件检测方面表现良好。     

MACSPMD:基于恶意API调用序列模式挖掘的恶意代码检测

基于动态分析的恶意代码检测方法由于能有效对抗恶意代码的多态和代码混淆技术,而且可以检测新的未知恶意代码等,因此得到了研究者的青睐。在这种情况下,恶意代码的编写者通过在恶意代码中嵌入大量反检测功能来逃避现有恶意代码动态检测方法的检测。针对该问题,提出了基于恶意API调用序列模式挖掘的恶意代码检测方法MACSPMD。首先,使用真机模拟恶意代码的实际运行环境来获取文件的动态API调用序列;其次,引入面向目标关联挖掘的概念,以挖掘出能够代表潜在恶意行为模式的恶意API调用序列模式;最后,将挖掘到的恶意API调用序列模式作为异常行为特征进行恶意代码的检测。基于真实数据集的实验结果表明,MACSPMD对未知和逃避型恶意代码进行检测的准确率分别达到了94.55%和97.73%,比其他基于API调用数据的恶意代码检测方法 的准确率分别提高了2.47%和2.66%,且挖掘过程消耗的时间更少。因此,MACSPMD能有效检测包括逃避型在内的已知和未知恶意代码。     

Malicious sequential pattern mining for automatic malware detection

Due to its damage to Internet security, malware (e.g., virus, worm, trojan) and its detection has caught the attention of both anti-malware industry and researchers for decades. To protect legitimate users from the attacks, the most significant line of defense against malware is anti-malware software products, which mainly use signature-based method for detection. However, this method fails to recognize new, unseen malicious executables. To solve this problem, in this paper, based on the instruction sequences extracted from the file sample set, we propose an effective sequence mining algorithm to discover malicious sequential patterns, and then All-Nearest-Neighbor (ANN) classifier is constructed for malware detection based on the discovered patterns. The developed data mining framework composed of the proposed sequential pattern mining method and ANN classifier can well characterize the malicious patterns from the collected file sample set to effectively detect newly unseen malware samples. A comprehensive experimental study on a real data collection is performed to evaluate our detection framework. Promising experimental results show that our framework outperforms other alternate data mining based detection methods in identifying new malicious executables.     

基于序列模式发现的恶意行为检测方法

为有效预防变形病毒和新出现的恶意软件,提出一种基于序列模式发现的恶意行为静态检测方法。将恶意代码转换为汇编代码,对其进行预处理,采用类Apriori算法完成序列模式发现,并去除正常模式,得到可用于未知恶意代码检测的模式集合。实验结果表明,该方法的正确率较高、漏报率较低。     

一种混合的Android恶意应用检测方法

针对静态检测和动态检测方式存在的问题,提出了一种基于混合方式的恶意移动应用检测方法。该方法采用静态分析和动态分析相结合的方式,通过静态分析获取权限特征和函数调用特征,通过动态分析在沙盒环境下借助于事件仿真获取系统调用序列并提取函数调用依赖关系特征;在此基础上,提出了一种基于集成学习的分类器构造方法,区分恶意应用和正常应用。在来自于第三方应用市场中的3000个样本集上进行了实验验证,结果表明基于混合方式的恶意应用检测效果要优于基于静态分析的方式和基于动态分析的方式;考虑多种类型特征的样本上的检测精度要高于采用单一特征刻画的样本上的值;采用集成分类器具有较好的检测精度。     

HDM-Analyser: a hybrid analysis approach based on data mining techniques for malware detection

Today’s security threats like malware are more sophisticated and targeted than ever, and they are growing at an unprecedented rate. To deal with them, various approaches are introduced. One of them is Signature-based detection, which is an effective method and widely used to detect malware; however, there is a substantial problem in detecting new instances. In other words, it is solely useful for the second malware attack. Due to the rapid proliferation of malware and the desperate need for human effort to extract some kinds of signature, this approach is a tedious solution; thus, an intelligent malware detection system is required to deal with new malware threats. Most of intelligent detection systems utilise some data mining techniques in order to distinguish malware from sane programs. One of the pivotal phases of these systems is extracting features from malware samples and benign ones in order to make at least a learning model. This phase is called “Malware Analysis” which plays a significant role in these systems. Since API call sequence is an effective feature for realising unknown malware, this paper is focused on extracting this feature from executable files. There are two major kinds of approach to analyse an executable file. The first type of analysis is “Static Analysis” which analyses a program in source code level. The second one is “Dynamic Analysis” that extracts features by observing program’s activities such as system requests during its execution time. Static analysis has to traverse the program’s execution path in order to find called APIs. Because it does not have sufficient information about decision making points in the given executable file, it is not able to extract the real sequence of called APIs. Although dynamic analysis does not have this drawback, it suffers from execution overhead. Thus, the feature extraction phase takes noticeable time. In this paper, a novel hybrid approach, HDM-Analyser, is presented which takes advantages of dynamic and static analysis methods for rising speed while preserving the accuracy in a reasonable level. HDM-Analyser is able to predict the majority of decision making points by utilising the statistical information which is gathered by dynamic analysis; therefore, there is no execution overhead. The main contribution of this paper is taking accuracy advantage of the dynamic analysis and incorporating it into static analysis in order to augment the accuracy of static analysis. In fact, the execution overhead has been tolerated in learning phase; thus, it does not impose on feature extraction phase which is performed in scanning operation. The experimental results demonstrate that HDM-Analyser attains better overall accuracy and time complexity than static and dynamic analysis methods.     

一种基于混合特征的移动终端恶意软件检测方法

随着移动终端恶意软件的种类和数量不断增大,本文针对Android系统恶意软件单特征检测不全面、误报率高等技术问题,提出一种基于动静混合特征的移动终端恶意软件检测方法,以提高检测的覆盖率、准确率和效率.该方法首先采用基于改进的CHI方法和凝聚层次聚类算法优化的K-Means方法构建高危权限和敏感API库,然后分别从静态分...     

基于机器学习算法的Android恶意程序检测系统

对于传统的恶意程序检测方法存在的缺点,针对将数据挖掘和机器学习算法被应用在未知恶意程序的检测方法进行研究。当前使用单一特征的机器学习算法无法充分发挥其数据处理能力,检测效果不佳。文中将语音识别模型与随机森林算法相结合,首次提出了综和APK文件多类特征统一建立N-gram模型,并应用随机森林算法用于未知恶意程序检测。首先,采用多种方式提取可以反映Android恶意程序行为的3类特征,包括敏感权限、DVM函数调用序列以及OpCodes特征;然后,针对每类特征建立N-gram模型,每个模型可以独立评判恶意程序行为;最后,3类特征模型统一加入随机森林算法进行学习,从而对Android程序进行检测。基于该方法实现了Android恶意程序检测系统,并对811个非恶意程序及826个恶意程序进行检测,准确率较高。综合各个评价指标,与其他相关工作对比,实验结果表明该系统在恶意程序检测准确率和有效性上表现更优。     

Classification of malware based on integrated static and dynamic features

Collection of dynamic information requires that malware be executed in a controlled environment; the malware unpacks itself as a preliminary to the execution process. On the other hand, while execution of malware is not needed in order to collect static information, the file must first be unpacked manually. None-the-less, if a file has been executed, it is possible to use both static and dynamic information in designing a single classification method.In this paper, we present the first classification method integrating static and dynamic features into a single test. Our approach improves on previous results based on individual features and reduces by half the time needed to test such features separately.Robustness to changes in malware development is tested by comparing results on two sets of malware, the first collected between 2003 and 2007, and the second collected between 2009 and 2010. When classifying the older set as compared to the entire data set, our integrated test demonstrates significantly more robustness than previous methods by losing just 2.7% in accuracy as opposed to a drop of 7%. We conclude that to achieve acceptable accuracy in classifying the latest malware, some older malware should be included in the set of data.     

DroidChain: A novel Android malware detection method based on behavior chains

The drastic increase of Android malware has led to strong interest in automating malware analysis. In this paper, to fight against malware variants and zero-day malware, we proposed DroidChain: a method combining static analysis and a behavior chain model. We transform the malware detection problem into more accessible matrix form. Using this method, we propose four kinds of malware models, including privacy leakage, SMS financial charges, malware installation, and privilege escalation. To reduce time complexity, we propose the WxShall-extend algorithm. We had moved the prototype to GitHub and evaluate using 1260 malware samples. Experimental malware detection results demonstrate accuracy, precision, and recall of 73%–93%, 71%–99%, and 42%–92%, respectively. Calculation time accounts for 6.58% of the well-known Warshall algorithm’s expense. Results demonstrate that our method, which can detect four kinds of malware simultaneously, is better than Androguard and Kirin.     

SBMDS: an interpretable string based malware detection system using SVM ensemble with bagging

Malicious executables are programs designed to infiltrate or damage a computer system without the owner’s consent, which have become a serious threat to the security of computer systems. There is an urgent need for effective techniques to detect polymorphic, metamorphic and previously unseen malicious executables of which detection fails in most of the commercial anti-virus software. In this paper, we develop interpretable string based malware detection system (SBMDS), which is based on interpretable string analysis and uses support vector machine (SVM) ensemble with Bagging to classify the file samples and predict the exact types of the malware. Interpretable strings contain both application programming interface (API) execution calls and important semantic strings reflecting an attacker’s intent and goal. Our SBMDS is carried out with four major steps: (1) first constructing the interpretable strings by developing a feature parser; (2) performing feature selection to select informative strings related to different types of malware; (3) followed by using SVM ensemble with bagging to construct the classifier; (4) and finally conducting the malware detector, which not only can detect whether a program is malicious or not, but also can predict the exact type of the malware. Our case study on the large collection of file samples collected by Kingsoft Anti-virus lab illustrate that: (1) The accuracy and efficiency of our SBMDS outperform several popular anti-virus software; (2) Based on the signatures of interpretable strings, our SBMDS outperforms data mining based detection systems which employ single SVM, Naive Bayes with bagging, Decision Trees with bagging; (3) Compared with the IMDS which utilizes the objective-oriented association (OOA) based classification on API calls, our SBMDS achieves better performance. Our SBMDS system has already been incorporated into the scanning tool of a commercial anti-virus software.