Error-Resilient Binary Neural Network Inference with Selective Recompute-Based Error Correction
Gil-Ho Kwak, Tae‐Hwan Kim · 2024
This paper presents an efficient error-correction technique aiming at error-resilient inference for binary neural networks. The proposed technique is designed to achieve error-resilience by recompute-based error correction against transient errors potentially encountered in inference systems. The recompute is performed selectively by exploiting the spatial locality in a feature map to minimize the overhead. For the CIFAR10 classification task, the BNN inference with the proposed technique achieves an accuracy of 80.97% at a bit-error rate of 3%, which is 41.48% higher than that achieved without error resilience. The compute overhead is 64.61% of that of the conventional triple recompute technique in terms of latency.