A Methodological Framework for Supporting the Operational Analysis of ML-Based Systems - Archive ouverte HAL
Communication Dans Un Congrès Année : 2024

A Methodological Framework for Supporting the Operational Analysis of ML-Based Systems

Afef Awadid
  • Fonction : Auteur
Kevin Mantissa
Xavier Leroux
  • Fonction : Auteur

Résumé

The engineering of Machine Learning (ML)-based systems is a com-plex task that involves integrating multiple disciplines and dealing with uncer-tainties inherent in ML techniques. To tackle this challenge, the confiance.ai re-search program1 has developed an end-to-end methodology for engineering ML-based systems. This methodology is based on existing standards and industrial practice and is supported by a web application called “the body-of-knowledge”. The end-to-end methodology provides guidelines covering all phases of the pro-cess of engineering ML-based systems. In this paper, we focus on the methodo-logical framework as part of this methodology and that is dedicated to the oper-ational analysis phase. This framework refers to a set of engineering activities captured in Capella models and guiding the operational analysis for Intended Pur-pose and Automation Objectives. This paper provides an overview of the pro-posed framework and illustrates its underlying steps with an example of an Au-tomated Driving System (ADS).
Fichier non déposé

Dates et versions

hal-04682052 , version 1 (30-08-2024)

Identifiants

  • HAL Id : hal-04682052 , version 1

Citer

Afef Awadid, Kevin Mantissa, Xavier Leroux, Boris Robert. A Methodological Framework for Supporting the Operational Analysis of ML-Based Systems. 31st International Conference on Systems Engineering (ICSEng), Oct 2024, Warsaw (Poland), Poland. ⟨hal-04682052⟩
53 Consultations
0 Téléchargements

Partager

More