mlsanityData to 7 Oct 2026

Assumption violations in causal discovery and the robustness of score matching

arXiv 2310.13387 ↗First posted 20 Oct 2023stat.ME cs.LGcited by 28

Not on the map. Why?

Abstract (arXiv)

When domain knowledge is limited and experimentation is restricted by ethical, financial, or time constraints, practitioners turn to observational causal discovery methods to recover the causal structure, exploiting the statistical properties of their data. Because causal discovery without further assumptions is an ill-posed problem, each algorithm comes with its own set of usually untestable assumptions, some of which are hard to meet in real datasets. Motivated by these considerations, this paper extensively benchmarks the empirical performance of recent causal discovery methods on observational i.i.d. data generated under different background conditions, allowing for violations of the critical assumptions required by each selected approach. Our experimental findings show that score matching-based methods demonstrate surprising performance in the false positive and false negative rate of the inferred graph in these challenging scenarios, and we provide theoretical insights into their performance. This work is also the first effort to benchmark the stability of causal discovery algorithms with respect to the values of their hyperparameters. Finally, we hope this paper will set a new standard for the evaluation of causal discovery methods and can serve as an accessible entry point for practitioners interested in the field, highlighting the empirical implications of different algorithm choices.

Key facts. Cited by 28 papers in the map (data through 7 Oct 2026). Builds on 20 of its 55 references that are in the map. First posted on arXiv 20 Oct 2023.

Builds on

20 of 55 references.

References · 20

  1. 2018DAGs with NO TEARS: Continuous Optimization for Structure Learningcited by 745
  2. 2013Causal Discovery with Continuous Additive Noise Modelscited by 252
  3. 2017Detecting causal associations in large nonlinear time series datasetscited by 240
  4. 2019Gradient-Based Neural DAG Learningcited by 227
  5. 2019Learning Sparse Nonparametric DAGscited by 207
  6. 2013CAM: Causal additive models, high-dimensional order search and penalized regressioncited by 175
  7. 2021Beware of the Simulated DAG! Causal Discovery Benchmarks May Be Easy To Gamecited by 159
  8. 2022Score matching enables causal discovery of nonlinear additive noise modelscited by 108
  9. 2017Causal Structure Learningcited by 84
  10. 2020High-recall causal discovery for autocorrelated time series with latent confounderscited by 82
  11. 2013Causal Networks: Semantics and Expressivenesscited by 67
  12. 2022Diffusion Models for Causal Discovery via Topological Orderingcited by 64
8 more
  1. 2017Invariant Causal Prediction for Sequential Datacited by 63
  2. 2017Gradient Estimators for Implicit Modelscited by 58
  3. 2023Scalable Causal Discovery with Score Matchingcited by 42
  4. 2023Causal Discovery with Score Matching on Additive Models with Arbitrary Noisecited by 41
  5. 2018Fast Conditional Independence Test for Vector Variables with Large Sample Sizescited by 16
  6. 2019Evaluating structure learning algorithms with a balanced scoring functioncited by 10
  7. 2017Causal Discovery in the Presence of Measurement Error: Identifiability Conditionscited by 9
  8. 2017Comparative Benchmarking of Causal Discovery Techniquescited by 6

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 1; 2024: 14; 2025: 4; 2026: 9.

28 papers.

Sort:

  1. 2024On the Challenges and Opportunities in Generative AIcited by 31
  2. 2025CausalRivers -- Scaling up benchmarking of causal discovery for real-world time-seriescited by 24
  3. 2024Signature Kernel Conditional Independence Tests in Causal Discovery for Stochastic Processescited by 18
  4. 2024The Landscape of Causal Discovery Data: Grounding Causal Discovery in Real-World Applicationscited by 17
  5. 2024Amortized Active Causal Induction with Deep Reinforcement Learningcited by 10
  6. 2024Demystifying amortized causal discovery with transformerscited by 10
  7. 2024Score matching through the roof: linear, nonlinear, and latent variables causal discoverycited by 8
  8. 2024A Skewness-Based Criterion for Addressing Heteroscedastic Noise in Causal Discoverycited by 7
  9. 2024Better Simulations for Validating Causal Discovery with the DAG-Adaptation of the Onion Methodcited by 6
  10. 2025Position: Causal Machine Learning Requires Rigorous Synthetic Experiments for Broader Adoptioncited by 6
  11. 2024Hybrid Top-Down Global Causal Discovery with Local Search for Linear and Nonlinear Additive Noise Modelscited by 5
  12. 2024Deriving Causal Order from Single-Variable Interventions: Guarantees & Algorithmcited by 5
16 more
  1. 2024Challenges and Considerations in the Evaluation of Bayesian Causal Discoverycited by 5
  2. 2024Scalable and Flexible Causal Discovery with an Efficient Test for Adjacencycited by 5
  3. 2023Shortcuts for causal discovery of nonlinear models by score matchingcited by 4
  4. 2025The Robustness of Differentiable Causal Discovery in Misspecified Scenarioscited by 4
  5. 2026Causal Representation Learning from General Environments under Nonparametric Mixingcited by 4
  6. 2024Efficient Differentiable Discovery of Causal Ordercited by 2
  7. 2024LoSAM: Local Search in Additive Noise Models with Mixed Mechanisms and General Noise for Global Causal Discoverycited by 1
  8. 2025Causal Ordering for Structure Learning from Time Seriescited by 1
  9. 2026Ordering-based Causal Discovery via Generalized Score Matchingcited by 1
  10. 2026Transformer Is Inherently a Causal Learnercited by 0
  11. 2026ReBeCA: Unveiling Interpretable Behavior Hierarchy behind the Iterative Self-Reflection of Language Models with Causal Analysiscited by 0
  12. 2026Adaptive-CaRe: Adaptive Causal Regularization for Robust Outcome Predictioncited by 0
  13. 2026Causal-Audit: A Framework for Risk Assessment of Assumption Violations in Time-Series Causal Discoverycited by 0
  14. 2026Causal Learning with the Invariance Principlecited by 0
  15. 2026NetCause: Counterfactual Learning for Root Cause Analysis in Large-Scale Networkscited by 0
  16. 2026How Does Bayesian Causal Discovery Fail? Characterising Structural Consequences in Linear Gaussian Networks under Latent Confoundingcited by 0