As “deep neural networks” (DNNs) achieve increasing accuracy, they are getting employed in increasingly diverse applications, including security-critical applications such as medical and defense. The worldwide revenue produced from the deployment of AI is expected to reach $190.6 billion by 2025. This immense use of DNNs has motivated the researchers to scrutinizingly study their hardware security vulnerability and propose countermeasures
Our paper presents a survey of 80+ techniques for the hardware security of DNNs. It reviews timing/memory/power/electromagnetic side-channel attacks, trojan attacks, fault-injection attacks, along with many defense techniques.
The paper is available here, accepted in Journal of Systems Architecture 2021.
Our paper presents a survey of 80+ techniques for the hardware security of DNNs. It reviews timing/memory/power/electromagnetic side-channel attacks, trojan attacks, fault-injection attacks, along with many defense techniques.
The paper is available here, accepted in Journal of Systems Architecture 2021.