Facilitating Change Implementation for Continuous ML-Safety Assurance

TL;DR

This paper introduces a method for integrating safety assurance into continuous machine learning deployment by linking safety argumentation with quantitative metrics, demonstrated on an autonomous vehicle system.

Contribution

It presents a novel approach to safety assurance that combines semantic safety argumentation with quantitative metrics for continuous ML system deployment.

Findings

Semantic tags enhance safety argumentation clarity

Quantitative metrics enable impact analysis of evidence

Application to autonomous vehicle braking system demonstrates practicality

Abstract

We propose a method for deploying a safety-critical machine-learning component into continuously evolving environments where an increased degree of automation in the engineering process is desired. We associate semantic tags with the safety case argumentation and turn each piece of evidence into a quantitative metric or a logic formula. With proper tool support, the impact can be characterized by a query over the safety argumentation tree to highlight evidence turning invalid. The concept is exemplified using a vision-based emergency braking system of an autonomous guided vehicle for factory automation.