Stealing Neural Network Structure Through Remote FPGA Side-Channel Analysis
- 19 August 2021
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Information Forensics and Security
- Vol. 16 (15566013) , 4377-4388
Abstract
Deep Neural Network (DNN) models have been extensively developed by companies for a wide range of applications. The development of a customized DNN model with great performance requires costly investments, and its structure (layers and hyper-parameters) is considered intellectual property and holds immense value. However, in this paper, we found the model secret is vulnerable when a cloud-based FPGA accelerator executes it. We demonstrate an end-to-end attack based on remote power side-channel analysis and machine-learning-based secret inference against different DNN models. The evaluation result shows that an attacker can reconstruct the layer and hyper-parameter sequence at over 90% accuracy using our method, which can significantly reduce her model development workload. We believe the threat presented by our attack is tangible, and new defense mechanisms should be developed against this threat.Keywords
Funding Information
- NSF (ECCS-2028269, DGE-2039634)
This publication has 45 references indexed in Scilit:
- Throughput-Optimized FPGA Accelerator for Deep Convolutional Neural NetworksACM Transactions on Reconfigurable Technology and Systems, 2017
- XGBoost2016
- Enabling FPGAs in Hyperscale Data Centers2015
- Enabling FPGAs in the cloud2014
- FPGAs in the Cloud: Booting Virtualized Hardware Accelerators with OpenStack2014
- Hey, you, get off of my cloud2009
- Multi-class AdaBoostStatistics and Its Interface, 2009
- Stochastic gradient boostingComputational Statistics & Data Analysis, 2002
- A survey of decision tree classifier methodologyIEEE Transactions on Systems, Man, and Cybernetics, 1991
- An introduction to computing with neural netsIEEE ASSP Magazine, 1987