Security Analysis of Deep Neural Networks Operating in the Presence of Cache Side-Channel Attacks

TL;DR

This paper analyzes the security risks of deep neural networks against cache side-channel attacks, demonstrating how attackers can reconstruct network architectures and proposing defenses to mitigate these vulnerabilities.

Contribution

It introduces DeepRecon, a novel cache side-channel attack for DNN architecture reconstruction, and evaluates defense strategies to improve model security.

Findings

Attack can accurately reconstruct VGG19 and ResNet50 architectures from a single inference.

Attackers can fingerprint model architecture and family using extracted attributes.

Proposed defenses can obfuscate attack observations and improve security.

Abstract

Recent work has introduced attacks that extract the architecture information of deep neural networks (DNN), as this knowledge enhances an adversary's capability to conduct black-box attacks against the model. This paper presents the first in-depth security analysis of DNN fingerprinting attacks that exploit cache side-channels. First, we define the threat model for these attacks: our adversary does not need the ability to query the victim model; instead, she runs a co-located process on the host machine victim's deep learning (DL) system is running and passively monitors the accesses of the target functions in the shared framework. Second, we introduce DeepRecon, an attack that reconstructs the architecture of the victim network by using the internal information extracted via Flush+Reload, a cache side-channel technique. Once the attacker observes function invocations that map directly to architecture attributes of the victim network, the attacker can reconstruct the victim's entire network architecture. In our evaluation, we demonstrate that an attacker can accurately reconstruct two complex networks (VGG19 and ResNet50) having observed only one forward propagation. Based on the extracted architecture attributes, we also demonstrate that an attacker can build a meta-model that accurately fingerprints the architecture and family of the pre-trained model in a transfer learning setting. From this meta-model, we evaluate the importance of the observed attributes in the fingerprinting process. Third, we propose and evaluate new framework-level defense techniques that obfuscate our attacker's observations. Our empirical security analysis represents a step toward understanding the DNNs' vulnerability to cache side-channel attacks.