One-Shot Multi-Label Causal Discovery in High-Dimensional Event Sequences

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Math, Hugo, Schön, Robin, Lienhart, Rainer
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866914155501453312
author Math, Hugo
Schön, Robin
Lienhart, Rainer
author_facet Math, Hugo
Schön, Robin
Lienhart, Rainer
contents Understanding causality in event sequences with thousands of sparse event types is critical in domains such as healthcare, cybersecurity, or vehicle diagnostics, yet current methods fail to scale. We present OSCAR, a one-shot causal autoregressive method that infers per-sequence Markov Boundaries using two pretrained Transformers as density estimators. This enables efficient, parallel causal discovery without costly global CI testing. On a real-world automotive dataset with 29,100 events and 474 labels, OSCAR recovers interpretable causal structures in minutes, while classical methods fail to scale, enabling practical scientific diagnostics at production scale.
format Preprint
id arxiv_https___arxiv_org_abs_2509_23213
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle One-Shot Multi-Label Causal Discovery in High-Dimensional Event Sequences
Math, Hugo
Schön, Robin
Lienhart, Rainer
Machine Learning
Artificial Intelligence
Understanding causality in event sequences with thousands of sparse event types is critical in domains such as healthcare, cybersecurity, or vehicle diagnostics, yet current methods fail to scale. We present OSCAR, a one-shot causal autoregressive method that infers per-sequence Markov Boundaries using two pretrained Transformers as density estimators. This enables efficient, parallel causal discovery without costly global CI testing. On a real-world automotive dataset with 29,100 events and 474 labels, OSCAR recovers interpretable causal structures in minutes, while classical methods fail to scale, enabling practical scientific diagnostics at production scale.
title One-Shot Multi-Label Causal Discovery in High-Dimensional Event Sequences
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2509.23213