Talks and presentations

FP-Rowhammer: Rowhammer-Based Device Fingerprinting

UC Davis, Davis, CA 95616 on Feb 16 2024
Device fingerprinting relies on attributes that capture heterogeneity in hardware and software configurations to extract unique and stable fingerprints. Fingerprinting countermeasures attempt to either present a uniform fingerprint across different devices through normalization or present different fingerprints for the same device each time through obfuscation. We present FP-Rowhammer, a Rowhammer-based device fingerprinting approach that can build unique and stable fingerprints even across devices with normalized or obfuscated hardware and software configurations. To this end, FP-Rowhammer leverages the DRAM manufacturing process variation that gives rise to unique distributions of Rowhammer-induced bit flips across different DRAM modules.

Aragorn: An Automated and Extensible Privacy-Enhancing System for Mobile Cameras

UC Davis, Davis, CA 95616 on Jan 21 2022
Mobile app developers often rely on high-quality cameras to implement rich functionalities. However, giving apps unfettered access to the mobile camera poses a serious threat to user privacy because camera frames may contain sensitive information that is beyond the scope of the app. To mitigate this privacy concern, we present Aragorn, a novel privacy-enhancing mobile camera system that automatically sanitizes camera frames to only contain user-authorized objects without needing any manual user input.

eXeGAN: Not all Malware is created Equal

Blue Hexagon, Sunnyvale, CA 94086 on Sep 20 2019
Antivirus engines use static analysis of executables to detect malware. In some cases, they prefer static analysis over dynamic analysis since dynamic analysis is time consuming and can be unreliable if the malware detects that it is being executed in a sandboxed environment for analysis. The advent of machine learning has greatly shifted malware detection in favor of these engines, since ML models tend to generalize, helping them detect completely new and unseen types of malware. We explore if malware authors can also leverage machine learning to evade detection by presenting eXeGAN, a generative adversarial network that mutates executables to evade ML and ruled based static detection, while preserving their functionality.