With machine learning techniques widely used to automate Android malware detection, it is important to investigate the robustness of these methods against evasion attacks. A recent work has proposed a novel problem-space attack on Android malware classifiers, where adversarial examples are generated by transforming Android malware samples while satisfying practical constraints. Aimed to address its limitations, we propose a new attack called EAGLE (Evasion Attacks Guided by Local Explanations), whose key idea is to leverage local explanations to guide the search for adversarial examples. We present a generic algorithmic framework for EAGLE attacks, which can be customized with specific feature increase and decrease operations to evade Android malware classifiers trained on different types of count features. We overcome practical challenges in implementing these operations for four different types of Android malware classifiers. Using two Android malware datasets, our results show that EAGLE attacks can be highly effective at finding functionable adversarial examples. We study the attack transferrability of malware variants created by EAGLE attacks across classifiers built with different classification models or trained on different types of count features. Our research further demonstrates that ensemble classifiers trained from multiple types of count features are not immune to EAGLE attacks. We also discuss possible defense mechanisms against EAGLE attacks.
more »
« less
An Evaluation of One-Class Feature Selection and Classification for Zero-Day Android Malware Detection
Security has become a serious problem for Android system as the number of Android malware increases rapidly. A great amount of effort has been devoted to protect Android devices against the threats of malware. Majority of the existing work use two-class classification methods which suffer the overfitting problem due to the lack of malicious samples. This will result in poor performance of detecting zero-day malware attacks. In this paper, we evaluated the performance of various one-class feature selection and classification methods for zero-day Android malware detection. Unlike two-class methods, one-class methods only use benign samples to build the detection model which overcomes the overfitting problem. Our results demonstrate the capability of the one-class methods over the two-class methods in detecting zero-day Android malware attacks.
more »
« less
- Award ID(s):
- 1757207
- PAR ID:
- 10156573
- Date Published:
- Journal Name:
- New generation
- ISSN:
- 1631-235X
- Page Range / eLocation ID:
- 105-111
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
The detection of zero-day attacks and vulnerabilities is a challenging problem. It is of utmost importance for network administrators to identify them with high accuracy. The higher the accuracy is, the more robust the defense mechanism will be. In an ideal scenario (i.e., 100% accuracy) the system can detect zero-day malware without being concerned about mistakenly tagging benign files as malware or enabling disruptive malicious code running as none-malicious ones. This paper investigates different machine learning algorithms to find out how well they can detect zero-day malware. Through the examination of 34 machine/deep learning classifiers, we found that the random forest classifier offered the best accuracy. The paper poses several research questions regarding the performance of machine and deep learning algorithms when detecting zero-day malware with zero rates for false positive and false negative.more » « less
-
Malicious software (malware) classification offers a unique challenge for continual learning (CL) regimes due to the volume of new samples received on a daily basis and the evolution of malware to exploit new vulnerabilities. On a typical day, antivirus vendors receive hundreds of thousands of unique pieces of software, both malicious and benign, and over the course of the lifetime of a malware classifier, more than a billion samples can easily accumulate. Given the scale of the problem, sequential training using continual learning techniques could provide substantial benefits in reducing training and storage overhead. To date, however, there has been no exploration of CL applied to malware classification tasks. In this paper, we study 11 CL techniques applied to three malware tasks covering common incremental learning scenarios, including task, class, and domain incremental learning (IL). Specifically, using two realistic, large-scale malware datasets, we evaluate the performance of the CL methods on both binary malware classification (Domain-IL) and multi-class malware family classification (Task-IL and Class-IL) tasks. To our surprise, continual learning methods significantly underperformed naive Joint replay of the training data in nearly all settings – in some cases reducing accuracy by more than 70 percentage points. A simple approach of selectively replaying 20% of the stored data achieves better performance, with 50% of the training time compared to Joint replay. Finally, we discuss potential reasons for the unexpectedly poor performance of the CL techniques, with the hope that it spurs further research on developing techniques that are more effective in the malware classification domain.more » « less
-
Android, the most dominant Operating System (OS), experiences immense popularity for smart devices for the last few years. Due to its' popularity and open characteristics, Android OS is becoming the tempting target of malicious apps which can cause serious security threat to financial institutions, businesses, and individuals. Traditional anti-malware systems do not suffice to combat newly created sophisticated malware. Hence, there is an increasing need for automatic malware detection solutions to reduce the risks of malicious activities. In recent years, machine learning algorithms have been showing promising results in classifying malware where most of the methods are shallow learners like Logistic Regression (LR). In this paper, we propose a deep learning framework, called Droid-NNet, for malware classification. However, our proposed method Droid-NNet is a deep learner that outperforms existing cutting-edge machine learning methods. We performed all the experiments on two datasets (Malgenome-215 & Drebin-215) of Android apps to evaluate Droid-NNet. The experimental result shows the robustness and effectiveness of Droid-NNet.more » « less
-
null (Ed.)Android is the most targeted mobile OS. Studies have found that repackaging is one of the most common techniques that adversaries use to distribute malware, and detecting such malware can be difficult because they share large parts of the code with benign apps. Other studies have highlighted the privacy implications of zero-permission sensors. In this work, we investigate if repackaged malicious apps utilize more sensors than the benign counterpart for malicious purposes. We analyzed 15,297 app pairs for sensor usage. We provide evidence that zero-permission sensors are indeed used by malicious apps to perform various activities. We use this information to train a robust classifier to detect repackaged malware in the wild.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- https://doi.org/10.1007/978-3-030-43020-7_15
