Revealing CNN Architectures via Side-Channel Analysis in Dataflow-based Inference Accelerators

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Weerasena, Hansika, Mishra, Prabhat
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910928514056192
author Weerasena, Hansika
Mishra, Prabhat
author_facet Weerasena, Hansika
Mishra, Prabhat
contents Convolutional Neural Networks (CNNs) are widely used in various domains, including image recognition, medical diagnosis and autonomous driving. Recent advances in dataflow-based CNN accelerators have enabled CNN inference in resource-constrained edge devices. These dataflow accelerators utilize inherent data reuse of convolution layers to process CNN models efficiently. Concealing the architecture of CNN models is critical for privacy and security. This article evaluates memory-based side-channel information to recover CNN architectures from dataflow-based CNN inference accelerators. The proposed attack exploits spatial and temporal data reuse of the dataflow mapping on CNN accelerators and architectural hints to recover the structure of CNN models. Experimental results demonstrate that our proposed side-channel attack can recover the structures of popular CNN models, namely, Lenet, Alexnet, VGGnet16, and YOLOv2.
format Preprint
id arxiv_https___arxiv_org_abs_2311_00579
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Revealing CNN Architectures via Side-Channel Analysis in Dataflow-based Inference Accelerators
Weerasena, Hansika
Mishra, Prabhat
Cryptography and Security
Hardware Architecture
Machine Learning
Convolutional Neural Networks (CNNs) are widely used in various domains, including image recognition, medical diagnosis and autonomous driving. Recent advances in dataflow-based CNN accelerators have enabled CNN inference in resource-constrained edge devices. These dataflow accelerators utilize inherent data reuse of convolution layers to process CNN models efficiently. Concealing the architecture of CNN models is critical for privacy and security. This article evaluates memory-based side-channel information to recover CNN architectures from dataflow-based CNN inference accelerators. The proposed attack exploits spatial and temporal data reuse of the dataflow mapping on CNN accelerators and architectural hints to recover the structure of CNN models. Experimental results demonstrate that our proposed side-channel attack can recover the structures of popular CNN models, namely, Lenet, Alexnet, VGGnet16, and YOLOv2.
title Revealing CNN Architectures via Side-Channel Analysis in Dataflow-based Inference Accelerators
topic Cryptography and Security
Hardware Architecture
Machine Learning
url https://arxiv.org/abs/2311.00579