Maat: Automatically Analyzing VirusTotal for Accurate Labeling and Effective Malware Detection

The malware analysis and detection research community relies on the online platform VirusTotal to label Android apps based on the scan results of around 60 antiviral scanners. Unfortunately, there are no standards on how to best interpret the scan results acquired from VirusTotal, which leads to the utilization of different threshold-based labeling strategies (e.g., if ten or more scanners deem an app malicious, it is considered malicious). While some of the utilized thresholds may be able to accurately approximate the ground truths of apps, the fact that VirusTotal changes the set and versions of the scanners it uses makes such thresholds unsustainable over time. We implemented a method, Maat, that tackles these issues of standardization and sustainability by automatically generating a Machine Learning (ML)-based labeling scheme, which outperforms threshold-based labeling strategies. Using the VirusTotal scan reports of 53K Android apps that span one year, we evaluated the applicability of Maat's ML-based labeling strategies by comparing their performance against threshold-based strategies. We found that such ML-based strategies (a) can accurately and consistently label apps based on their VirusTotal scan reports, and (b) contribute to training ML-based detection methods that are more effective at classifying out-of-sample apps than their threshold-based counterparts.

Towards Accurate Labeling of Android Apps for Reliable Malware Detection

In training their newly-developed malware detection methods, researchers rely on threshold-based labeling strategies that interpret the scan reports provided by online platforms, such as VirusTotal. The dynamicity of this platform renders those labeling strategies unsustainable over prolonged periods, which leads to inaccurate labels. Using inaccurately labeled apps to train and evaluate malware detection methods significantly undermines the reliability of their results, leading to either dismissing otherwise promising detection approaches or adopting intrinsically inadequate ones. The infeasibility of generating accurate labels via manual analysis and the lack of reliable alternatives force researchers to utilize VirusTotal to label apps. In the paper, we tackle this issue in two manners. Firstly, we reveal the aspects of VirusTotal's dynamicity and how they impact threshold-based labeling strategies and provide actionable insights on how to use these labeling strategies given VirusTotal's dynamicity reliably. Secondly, we motivate the implementation of alternative platforms by (a) identifying VirusTotal limitations that such platforms should avoid, and (b) proposing an architecture of how such platforms can be constructed to mitigate VirusTotal's limitations.