Tracing Distribution Shifts with Causal System Maps

TL;DR

This paper introduces ML System Maps, a causal modeling approach that visualizes and attributes distribution shifts in ML systems, aiming to improve root-cause analysis and system monitoring.

Contribution

It proposes a novel causal mapping framework with layered views to systematically trace and attribute causes of distribution shifts in ML systems.

Findings

Conceptual framework for causal system maps

Layered visualization of propagation paths

Research agenda for development and evaluation

Abstract

Monitoring machine learning (ML) systems is hard, with standard practice focusing on detecting distribution shifts rather than their causes. Root-cause analysis often relies on manual tracing to determine whether a shift is caused by software faults, data-quality issues, or natural change. We propose ML System Maps -- causal maps that, through layered views, make explicit the propagation paths between the environment and the ML system's internals, enabling systematic attribution of distribution shifts. We outline the approach and a research agenda for its development and evaluation.