CausalOps -- Towards an Industrial Lifecycle for Causal Probabilistic Graphical Models

TL;DR

The paper introduces CausalOps, a lifecycle framework for causal probabilistic graphical models, designed to scale causal engineering in industry by drawing on DevOps and MLOps practices. It formalizes a seven-facet lifecycle (Arrange, Create, Test, Publish, Operate, Monitor, Document), defines boundary objects and a set of participating entities, and outlines artifacts that accompany each facet. The authors provide conceptual integration scenarios in automotive safety (SOTIF/ISO 26262) and manufacturing to show how CausalOps can augment existing processes and safety Argumentation. They discuss the interdisciplinary requirements, boundary-object theory, and tooling implications, and highlight challenges such as automated verification and organizational adoption as directions for future work. Overall, CausalOps aims to bridge theory and practice, offering a roadmap for industrializing causal engineering across domains.

Abstract

Causal probabilistic graph-based models have gained widespread utility, enabling the modeling of cause-and-effect relationships across diverse domains. With their rising adoption in new areas, such as automotive system safety and machine learning, the need for an integrated lifecycle framework akin to DevOps and MLOps has emerged. Currently, a process reference for organizations interested in employing causal engineering is missing. To address this gap and foster widespread industrial adoption, we propose CausalOps, a novel lifecycle framework for causal model development and application. By defining key entities, dependencies, and intermediate artifacts generated during causal engineering, we establish a consistent vocabulary and workflow model. This work contextualizes causal model usage across different stages and stakeholders, outlining a holistic view of creating and maintaining them. CausalOps' aim is to drive the adoption of causal methods in practical applications within interested organizations and the causality community.

Figures (6)

• Figure 1: Overview of various application-relevant properties of causal graphs, relations, and models
• Figure 2: High-level concept of the seven lifecycle facets of CausalOps
• Figure 3: Disciplines influencing the theoretical foundation of CausalOps
• Figure 4: Overview of relevant entities, their interdependencies, exemplary related artifacts, and their contribution in the individual lifecycle facets
• Figure 5: Overview of artifacts associated with individual lifecycle facets and their dissemination throughout CausalOps