A Practical Introduction to Side-Channel Extraction of Deep Neural Network Parameters

被引:6
作者
Joud, Raphael [1 ,2 ]
Moellic, Pierre-Alain [1 ,2 ]
Pontie, Simon [1 ,2 ]
Rigaud, Jean-Baptiste [3 ]
机构
[1] CEA Tech, Ctr CMP, Equipe Commune CEA Tech Mines St Etienne, F-13541 Gardanne, France
[2] Univ Grenoble Alpes, Leti, CEA, F-38000 Grenoble, France
[3] CEA, Mines St Etienne, Leti, Ctr CMP, F-13541 Gardanne, France
来源
SMART CARD RESEARCH AND ADVANCED APPLICATIONS, CARDIS 2022 | 2023年 / 13820卷
关键词
Side-channel analysis; Confidentiality; Machine Learning; Neural network;
D O I
10.1007/978-3-031-25319-5_3
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Model extraction is a major threat for embedded deep neural network models that leverages an extended attack surface. Indeed, by physically accessing a device, an adversary may exploit side-channel leakages to extract critical information of a model (i.e., its architecture or internal parameters). Different adversarial objectives are possible including a fidelity-based scenario where the architecture and parameters are precisely extracted (model cloning). We focus this work on software implementation of deep neural networks embedded in a high-end 32-bit microcontroller (Cortex-M7) and expose several challenges related to fidelity-based parameters extraction through side-channel analysis, from the basic multiplication operation to the feed-forward connection through the layers. To precisely extract the value of parameters represented in the single-precision floating point IEEE-754 standard, we propose an iterative process that is evaluated with both simulations and traces from a Cortex-M7 target. To our knowledge, this work is the first to target such an high-end 32-bit platform. Importantly, we raise and discuss the remaining challenges for the complete extraction of a deep neural network model, more particularly the critical case of biases.
引用
收藏
页码:45 / 65
页数:21
相关论文
共 15 条
[1]  
Batina L, 2019, PROCEEDINGS OF THE 28TH USENIX SECURITY SYMPOSIUM, P515
[2]   SNIFF: Reverse Engineering of Neural Networks With Fault Attacks [J].
Breier, Jakub ;
Jap, Dirmanto ;
Hou, Xiaolu ;
Bhasin, Shivam ;
Liu, Yang .
IEEE TRANSACTIONS ON RELIABILITY, 2022, 71 (04) :1527-1539
[3]   Cryptanalytic Extraction of Neural Network Models [J].
Carlini, Nicholas ;
Jagielski, Matthew ;
Mironov, Ilya .
ADVANCES IN CRYPTOLOGY - CRYPTO 2020, PT III, 2020, 12172 :189-218
[4]   Side channel attacks for architecture extraction of neural networks [J].
Chabanne, Herve ;
Danger, Jean-Luc ;
Guiga, Linda ;
Kuhne, Ulrich .
CAAI TRANSACTIONS ON INTELLIGENCE TECHNOLOGY, 2021, 6 (01) :3-16
[5]  
Dubey A, 2020, PROCEEDINGS OF THE 2020 IEEE INTERNATIONAL SYMPOSIUM ON HARDWARE ORIENTED SECURITY AND TRUST (HOST), P197, DOI [10.1109/HOST45689.2020.9300276, 10.1109/host45689.2020.9300276]
[6]  
Dumont Mathieu, 2021, 2021 IEEE 7th World Forum on Internet of Things (WF-IoT), P616, DOI 10.1109/WF-IoT51360.2021.9595075
[7]  
Gongye C., 2020, Design Automation Conference, P1
[8]   Reverse Engineering Convolutional Neural Networks Through Side-channel Information Leaks [J].
Hua, Weizhe ;
Zhang, Zhiru ;
Suh, G. Edward .
2018 55TH ACM/ESDA/IEEE DESIGN AUTOMATION CONFERENCE (DAC), 2018,
[9]  
Jagielski M, 2020, PROCEEDINGS OF THE 29TH USENIX SECURITY SYMPOSIUM, P1345
[10]  
Joud R., 2021, 2021 IEEE 7 WORLD FO