Causal Models in Requirement Specifications for Machine Learning: A vision

TL;DR

This paper advocates for using causal modeling in requirements engineering to systematically incorporate domain knowledge into ML system design, demonstrated through an industrial fault detection case.

Contribution

It introduces a workflow for eliciting data and model requirements from high-level knowledge using causal models in requirements engineering.

Findings

Demonstrated on an industrial fault detection system

Proposed workflow for integrating causal models into RE

Outlined future research directions for causal modeling in RE

Abstract

Specifying data requirements for machine learning (ML) software systems remains a challenge in requirements engineering (RE). This vision paper explores causal modelling as an RE activity that allows the systematic integration of prior domain knowledge into the design of ML software systems. We propose a workflow to elicit low-level model and data requirements from high-level prior knowledge using causal models. The approach is demonstrated on an industrial fault detection system. This paper outlines future research needed to establish causal modelling as an RE practice.