An Energy-Efficient Neural Network Accelerator with Improved Protections Against Fault-Attacks
- Resource Type
- Conference
- Authors
- Maji, Saurav; Lee, Kyungmi; Gongye, Cheng; Fei, Yunsi; Chandrakasan, Anantha P.
- Source
- ESSCIRC 2023- IEEE 49th European Solid State Circuits Conference (ESSCIRC) Solid State Circuits Conference (ESSCIRC), ESSCIRC 2023- IEEE 49th European. :233-236 Sep, 2023
- Subject
- Communication, Networking and Broadcast Technologies
Components, Circuits, Devices and Systems
Computing and Processing
Photonics and Electrooptics
Signal Processing and Analysis
Degradation
Computational modeling
Fault detection
Europe
Artificial neural networks
Electrical fault detection
Energy efficiency
- Language
- ISSN
- 2643-1319
Embedded neural network (NN) implementations are susceptible to misclassification under fault attacks. Injecting strong electromagnetic (EM) pulses is a non-invasive yet detrimental attack that affects the NN operations by (i) causing faults in the NN model/inputs while being read by the NN computation unit, and (ii) corrupting NN computations results to cause misclassification eventually. This paper presents the first ASIC demonstration of an energy-efficient NN accelerator with inbuilt fault attack detection. We incorporated lightweight cryptography-aided checks using the Craft cipher for on-chip verification to detect model/input errors and also as a fault detection sensor. Our developed ASIC has demonstrated excellent error detection capabilities (100% detection for 100k error attempts) with a minimal area overhead of 5.9% and negligible NN accuracy degradation.