Hojjat Aghakhani (University of California, Santa Barbara), Fabio Gritti (University of California, Santa Barbara), Francesco Mecca (Università degli Studi di Torino), Martina Lindorfer (TU Wien), Stefano Ortolani (Lastline Inc.), Davide Balzarotti (Eurecom), Giovanni Vigna (University of California, Santa Barbara), Christopher Kruegel (University of California, Santa Barbara)

Machine learning techniques are widely used in addition to signatures and heuristics to increase the detection rate of anti-malware software, as they automate the creation of detection models, making it possible to handle an ever-increasing number of new malware samples. In order to foil the analysis of anti-malware systems and evade detection, malware uses packing and other forms of obfuscation. However, few realize that benign applications use packing and obfuscation as well, to protect intellectual property and prevent license abuse.

In this paper, we study how machine learning based on static analysis features operates on packed samples. Malware researchers have often assumed that packing would prevent machine learning techniques from building effective classifiers. However, both industry and academia have published results that show that machine-learning-based classifiers can achieve good detection rates, leading many experts to think that classifiers are simply detecting the fact that a sample is packed, as packing is more prevalent in malicious samples. We show that, different from what is commonly assumed, packers do preserve some information when packing programs that is “useful” for malware classification. However, this information does not necessarily capture the sample’s behavior. We demonstrate that the signals extracted from packed executables are not rich enough for machine-learning-based models to (1) generalize their knowledge to operate on unseen packers, and (2) be robust against adversarial examples. We also show that a naïve application of machine learning techniques results in a substantial number of false positives, which, in turn, might have resulted in incorrect labeling of ground-truth data used in past work.

View More Papers

SPEECHMINER: A Framework for Investigating and Measuring Speculative Execution...

Yuan Xiao (The Ohio State University), Yinqian Zhang (The Ohio State University), Radu Teodorescu (The Ohio State University)

Read More

Revisiting Leakage Abuse Attacks

Laura Blackstone (Brown University), Seny Kamara (Brown University), Tarik Moataz (Brown University)

Read More

Measuring the Deployment of Network Censorship Filters at Global...

Ram Sundara Raman (University of Michigan), Adrian Stoll (University of Michigan), Jakub Dalek (Citizen Lab, University of Toronto), Reethika Ramesh (University of Michigan), Will Scott (Independent), Roya Ensafi (University of Michigan)

Read More

Decentralized Control: A Case Study of Russia

Reethika Ramesh (University of Michigan), Ram Sundara Raman (University of Michgan), Matthew Bernhard (University of Michigan), Victor Ongkowijaya (University of Michigan), Leonid Evdokimov (Independent), Anne Edmundson (Independent), Steven Sprecher (University of Michigan), Muhammad Ikram (Macquarie University), Roya Ensafi (University of Michigan)

Read More