mlsanityData to 14 Oct 2026

Falsifying Causal Graphs With Outlier Events

arXiv 2607.12145 ↗First posted 13 Jul 2026stat.ML cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

True causal relationships are rarely known, and inferring causal graphs from data is hard. A fundamental challenge is how to assess whether a given causal graph is good in the absence of a ground truth. We propose falsifying candidate causal graphs based on whether they can explain the propagation of an outlier event. Our approach leverages a key principle: weak outliers rarely cause strong ones. While this principle has previously been used in root cause analysis to identify root causes without prior knowledge of the graph, we turn it on its head and use it to falsify candidate causal graphs whose implied outlier propagation is inconsistent with the data. To this end, we present the first statistical tests for the hypothesis that a candidate graph is the true causal graph, and show they have false positive control, power guarantees against incorrect causal graphs, and can operate with a single outlier sample.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 36 of its 83 references that are in the map. First posted on arXiv 13 Jul 2026.

Builds on

36 of 83 references.

References · 36

  1. 2018DAGs with NO TEARS: Continuous Optimization for Structure Learningcited by 752
  2. 2015Causal inference using invariant prediction: identification and confidence intervalscited by 536
  3. 2020Differentiable Causal Discovery from Interventional Datacited by 218
  4. 2019Causal Discovery from Heterogeneous/Nonstationary Data with Independent Changescited by 166
  5. 2017Invariant Causal Prediction for Nonlinear Modelscited by 140
  6. 2023Causal-learn: Causal Discovery in Pythoncited by 109
  7. 2018Characterizing and Learning Equivalence Classes of Causal DAGs under Interventionscited by 76
  8. 2019Causal structure based root cause analysis of outlierscited by 65
  9. 2022Greedy Relaxations of the Sparsest Permutation Algorithmcited by 64
  10. 2015Learning Causal Graphs with Small Interventionscited by 63
  11. 2017Invariant Causal Prediction for Sequential Datacited by 63
  12. 2022Causal Inference-Based Root Cause Analysis for Online Service Systems with Intervention Recognitioncited by 56
24 more
  1. 2022Causal Discovery in Heterogeneous Environments Under the Sparse Mechanism Shift Hypothesiscited by 55
  2. 2013Causal Discovery from Changescited by 52
  3. 2022Causal de Finetti: On the Identification of Invariant Causal Structure in Exchangeable Datacited by 42
  4. 2022DoWhy-GCM: An extension of DoWhy for causal inference in graphical causal modelscited by 41
  5. 2018Learning and Testing Causal Models with Interventionscited by 29
  6. 2016Testing Bayesian Networkscited by 23
  7. 2023Self-Compatibility: Evaluating Causal Discovery without Ground Truthcited by 22
  8. 2019Causal discovery in heavy-tailed modelscited by 20
  9. 2024Root Cause Analysis of Outliers with Missing Structural Knowledgecited by 19
  10. 2022Verification and search algorithms for causal DAGscited by 18
  11. 2023textttcausalAssembly: Generating Realistic Production Data for Benchmarking Causal Discoverycited by 18
  12. 2017Choosing Between Methods of Combining p-valuescited by 17
  13. 2023The PetShop Dataset -- Finding Causes of Performance Issues across Microservicescited by 14
  14. 2023Toward Falsifying Causal Graphs Using a Permutation-Based Testcited by 11
  15. 2024Root cause discovery via permutations and Cholesky decompositioncited by 11
  16. 2016Estimating and Controlling the False Discovery Rate for the PC Algorithm Using Edge-Specific P-Valuescited by 8
  17. 2020Evaluation of Causal Structure Learning Algorithms via Risk Estimationcited by 8
  18. 2022On Testability of the Front-Door Model via Verma Constraintscited by 8
  19. 2023Goodness-of-Fit Tests for Linear Non-Gaussian Structural Equation Modelscited by 4
  20. 2025Root Cause Analysis of Outliers in Unknown Cyclic Graphscited by 4
  21. 2024Cross-validating causal discovery via Leave-One-Variable-Outcited by 3
  22. 2025On Different Notions of Redundancy in Conditional-Independence-Based Discovery of Graphical Modelscited by 3
  23. 2025Toward Universal Laws of Outlier Propagationcited by 3
  24. 2025From Guess2Graph: When and How Can Unreliable Experts Safely Boost Causal Discovery in Finite Samples?cited by 3

Built on by

Not cited yet.