MLSMM: Machine Learning Security Maturity Model

TL;DR

This paper introduces MLSMM, a framework to evaluate and improve security practices in ML software development, aiming to foster collaboration between industry and academia.

Contribution

It proposes the first ML security maturity model with structured levels across the ML development lifecycle, filling a gap in security assessment tools.

Findings

Initial model outlining security practice levels

Framework organized along ML development stages

Encourages industry-academia collaboration

Abstract

Assessing the maturity of security practices during the development of Machine Learning (ML) based software components has not gotten as much attention as traditional software development. In this Blue Sky idea paper, we propose an initial Machine Learning Security Maturity Model (MLSMM) which organizes security practices along the ML-development lifecycle and, for each, establishes three levels of maturity. We envision MLSMM as a step towards closer collaboration between industry and academia.