Triggering Stealthy Feature Map Backdoors via Physical Fault Injection in Embedded Neural Networks
For security researchers and practitioners of embedded AI systems, this paper reveals a novel attack vector at the intersection of hardware and algorithmic vulnerabilities, stressing the need for cross-level defenses.
This work introduces a cross-level attack that bridges physical fault injection with algorithmic backdoor attacks on embedded neural networks, demonstrating that physically induced intermediate perturbations can serve as stealthy triggers. The attack is shown to be effective on an ARM Cortex-M4 microcontroller, evading existing backdoor defenses that assume input-space triggers.
Fault injection (FI) attacks on embedded neural network (NN) implementations primarily focus on inducing misclassification by corrupting weights or intermediate computations, overlooking their interaction with algorithmic adversarial threats. In this work, we present a cross-level attack that bridges implementation-level physical faults to algorithm-level adversarial attacks. By characterizing fault-induced data perturbations during NN inference, we connect FI with backdoor learning, enabling system-level attacks that jointly exploit implementation- and algorithm-level vulnerabilities. Specifically, we propose a precise fault-injection method that reliably manipulates targeted register values to tractable states during execution. Leveraging this level of FI precision, we propose a novel end-to-end feature map-level backdoor attack, where physically induced intermediate perturbations serve as stealthy triggers. Unlike conventional input-based backdoors, our trigger is activated only under physical faults, causing the NN to exhibit adversarial behavior that compromises system integrity while remaining benign during normal operation. We demonstrate that such physically triggered backdoors can be mounted on embedded NN platforms and remain effective against existing backdoor defenses that typically assume input-space triggers. We showcase the attack practicality using electromagnetic FI on convolutional neural networks implemented on ARM Cortex-M4 microcontroller, which is a common platform for constrained embedded applications. Our results highlight a novel attack vector at the intersection of hardware and algorithmic levels, stressing the need for defenses across abstraction levels.