Concept Enforcement and Modularization as Methods for the ISO 26262 Safety Argumentation of Neural Networks

Gesina Schwalbe (1, 2) , Martin Schels (2)

1 University of Bamberg
2 Continental Automotive GmbH

Neural networks (NN) are prone to systematic faults which are hard to detect using the methods recommended by the ISO 26262 automotive functional safety standard. In this paper we propose a unified approach to two methods for NN safety argumentation: Assignment of human interpretable concepts to the internal representation of NNs to enable modularization and formal verification. Feasibility of the required concept embedding analysis is demonstrated in a minimal example and important aspects for generalization are investigated. The contribution of the methods is derived from a proposed generic argumentation structure for a NN model safety case.

Mots clés: concept enforcement; machine learning; neural networks; functional safety; ISO 26262; goal structuring notation; explainable AI

Domaines: Vision par ordinateur et reconnaissance de formes [cs.CV]; Apprentissage [cs.LG]

Gesina Schwalbe, Martin Schels. Concept Enforcement and Modularization as Methods for the ISO 26262 Safety Argumentation of Neural Networks. 10th European Congress on Embedded Real Time Software and Systems (ERTS 2020), Jan 2020, Toulouse, France. ⟨hal-02442796⟩