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.

Leest, J., Gerostathopoulos, I., Raibulet, C., Lago, P. (2026). Explaining Shifts in Machine Learning Systems with Causal Maps. In CAIN '26: Proceedings of the IEEE/ACM 5th International Conference on AI Engineering - Software Engineering for AI (pp.180-185) [10.1145/3793653.3793782].

Explaining Shifts in Machine Learning Systems with Causal Maps

Raibulet C.;
2026

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.
paper
causality; distribution shift; machine learning; root-cause analysis;
English
5th IEEE International Conference on AI Engineering - Software Engineering for AI, CAIN 2026 - April 12 - 13, 2026
2026
CAIN '26: Proceedings of the IEEE/ACM 5th International Conference on AI Engineering - Software Engineering for AI
9798400724756
2026
180
185
open
Leest, J., Gerostathopoulos, I., Raibulet, C., Lago, P. (2026). Explaining Shifts in Machine Learning Systems with Causal Maps. In CAIN '26: Proceedings of the IEEE/ACM 5th International Conference on AI Engineering - Software Engineering for AI (pp.180-185) [10.1145/3793653.3793782].
File in questo prodotto:
File Dimensione Formato  
Leest et al-2026-IEEE/ACM-VoR.pdf

accesso aperto

Tipologia di allegato: Publisher’s Version (Version of Record, VoR)
Licenza: Creative Commons
Dimensione 549.47 kB
Formato Adobe PDF
549.47 kB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/10281/630682
Citazioni
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
Social impact