Tracing Distribution Shifts with Causal System Maps

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Leest, Joran, Gerostathopoulos, Ilias, Lago, Patricia, Raibulet, Claudia
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918173337452544
author Leest, Joran
Gerostathopoulos, Ilias
Lago, Patricia
Raibulet, Claudia
author_facet Leest, Joran
Gerostathopoulos, Ilias
Lago, Patricia
Raibulet, Claudia
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2510_23528
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Tracing Distribution Shifts with Causal System Maps
Leest, Joran
Gerostathopoulos, Ilias
Lago, Patricia
Raibulet, Claudia
Software Engineering
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.
title Tracing Distribution Shifts with Causal System Maps
topic Software Engineering
url https://arxiv.org/abs/2510.23528