Fast and Accurate Error Simulation for CNNs against Soft Errors

TL;DR

This paper introduces a fast and accurate error simulation framework for CNNs to assess robustness against soft errors, bridging fault injection and error simulation with high accuracy and significant speedup.

Contribution

It presents a novel error simulation engine for CNNs that uses validated error models, achieving high accuracy and speed compared to existing fault injection tools.

Findings

Achieves about 99% accuracy compared to SASSIFI.

Provides a 44x to 63x speedup over TensorFI.

Effectively models fault effects with validated error patterns.

Abstract

The great quest for adopting AI-based computation for safety-/mission-critical applications motivates the interest towards methods for assessing the robustness of the application w.r.t. not only its training/tuning but also errors due to faults, in particular soft errors, affecting the underlying hardware. Two strategies exist: architecture-level fault injection and application-level functional error simulation. We present a framework for the reliability analysis of Convolutional Neural Networks (CNNs) via an error simulation engine that exploits a set of validated error models extracted from a detailed fault injection campaign. These error models are defined based on the corruption patterns of the output of the CNN operators induced by faults and bridge the gap between fault injection and error simulation, exploiting the advantages of both approaches. We compared our methodology against SASSIFI for the accuracy of functional error simulation w.r.t. fault injection, and against TensorFI in terms of speedup for the error simulation strategy. Experimental results show that our methodology achieves about 99\% accuracy of the fault effects w.r.t. SASSIFI, and a speedup ranging from 44x up to 63x w.r.t. TensorFI, that only implements a limited set of error models.