MILR: Mathematically Induced Layer Recovery for Plaintext Space Error Correction of CNNs

TL;DR

This paper introduces MILR, a software-based method for CNN error correction that exploits mathematical relationships within the network to enable self-healing from bit errors, enhancing robustness without hardware changes.

Contribution

MILR is a novel error correction system for CNNs that uses mathematical layer relationships to recover from weight and layer errors without hardware modifications.

Findings

MILR can correct single and multi-bit errors in CNN weights.

MILR enables self-healing of CNNs without hardware or network modifications.

The method is suitable for plaintext-space error correction in CNNs.

Abstract

The increased use of Convolutional Neural Networks (CNN) in mission critical systems has increased the need for robust and resilient networks in the face of both naturally occurring faults as well as security attacks. The lack of robustness and resiliency can lead to unreliable inference results. Current methods that address CNN robustness require hardware modification, network modification, or network duplication. This paper proposes MILR a software based CNN error detection and error correction system that enables self-healing of the network from single and multi bit errors. The self-healing capabilities are based on mathematical relationships between the inputs,outputs, and parameters(weights) of a layers, exploiting these relationships allow the recovery of erroneous parameters (weights) throughout a layer and the network. MILR is suitable for plaintext-space error correction (PSEC) given its ability to correct whole-weight and even whole-layer errors in CNNs.