Toward Certification of Machine-Learning Systems for Low Criticality Airborne Applications

TL;DR

This paper examines how existing aviation certification standards can be adapted for low-criticality machine learning systems, proposing assumptions to align ML development with certification requirements.

Contribution

It analyzes current standards and demonstrates that low-criticality ML systems can meet certification objectives under specific development assumptions.

Findings

Certification objectives achievable for low-criticality ML systems

Identification of assumptions to align ML development with standards

Analysis of incompatibilities between traditional standards and ML technology

Abstract

The exceptional progress in the field of machine learning (ML) in recent years has attracted a lot of interest in using this technology in aviation. Possible airborne applications of ML include safety-critical functions, which must be developed in compliance with rigorous certification standards of the aviation industry. Current certification standards for the aviation industry were developed prior to the ML renaissance without taking specifics of ML technology into account. There are some fundamental incompatibilities between traditional design assurance approaches and certain aspects of ML-based systems. In this paper, we analyze the current airborne certification standards and show that all objectives of the standards can be achieved for a low-criticality ML-based system if certain assumptions about ML development workflow are applied.