中文 English
Home
TECHNICAL PAPERS

Vulnerability of Neural Networks Deployed As Black Boxes Across Accelerated HW Through Electromagnetic Side Channels

popularity

This technical paper titled “Can one hear the shape of a neural network?: Snooping the GPU via Magnetic Side Channel” was presented by researchers at Columbia University, Adobe Research and University of Toronto at the 31st USENIX Security Symposium in August 2022.

Abstract:
“Neural network applications have become popular in both enterprise and personal settings. Network solutions are tuned meticulously for each task, and designs that can robustly resolve queries end up in high demand. As the commercial value of accurate and performant machine learning models increases, so too does the demand to protect neural architectures as confidential investments. We explore the vulnerability of neural networks deployed as black boxes across accelerated hardware through electromagnetic side channels.

We examine the magnetic flux emanating from a graphics processing unit’s power cable, as acquired by a cheap $3 induction sensor, and find that this signal betrays the detailed topology and hyperparameters of a black-box neural network model. The attack acquires the magnetic signal for one query with unknown input values, but known input dimension and batch size. The network reconstruction is possible due to the modular layer sequence in which deep neural networks are evaluated. We find that each layer component’s evaluation produces an identifiable magnetic signal signature, from which layer topology, width, function type, and sequence order can be inferred using a suitably trained classifier and a joint consistency optimization based on integer programming.

We study the extent to which network specifications can be recovered, and consider metrics for comparing network similarity. We demonstrate the potential accuracy of this side channel attack in recovering the details for a broad range of network architectures, including random designs. We consider applications that may exploit this novel side channel exposure, such as adversarial transfer attacks. In response, we discuss countermeasures to protect against our method and other similar snooping techniques.”

Find the prepublication technical paper here and the slides here.

Authors:
Henrique Teles Maia and Chang Xiao, Columbia University; Dingzeyu Li, Adobe Research; Eitan Grinspun, Columbia University & University of Toronto; Changxi Zheng, Columbia University.

Further Security Reading
Chip Backdoors: Assessing The Threat
Steps are being taken to minimize problems, but they will take years to implement.
Security Risks Widen With Commercial Chiplets
Choosing components from a multi-vendor menu holds huge promise for reducing costs and time-to-market, but it’s not as simple as it sounds.
Securing Heterogeneous Integration At The Chiplet, Interposer, And System-In-Package Levels (FICS-University Of Florida)
Verifying Side-Channel Security Pre-Silicon
Complexity and new applications are pushing security much further to the left in the design flow.



Leave a Reply


(Note: This name will be displayed publicly)