Pith. sign in

REVIEW 2 major objections 4 minor 39 references

Knowing which stage each variable comes from lets causal discovery recover multistage process graphs more accurately than methods that ignore that structure.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

OCDM recovers multistage causal DAGs by stage-constrained score-based order search plus STG-NN pruning, outperforming baselines on layered synthetic and causalAssembly data.

T0 review reviewed 2026-07-11 challenge →

load-bearing objection Solid engineering fix for multistage causal discovery: stage-constrained DiffAN order search plus STG-NN pruning beats baselines on layered graphs and causalAssembly, with the usual additive-noise caveats. the 2 major comments →

arxiv 2607.03971 v1 pith:DWTUV2K3 submitted 2026-07-04 cs.LG cs.AI

Order-based Causal Discovery for Multistage Processes

Eun-Yeol Ma , Junsub Jung , Heeyoung Kim This is my paper
classification cs.LG cs.AI
keywords causal discoverymultistage processescausal order searchlayered DAGsscore matchingstochastic gated neural networkspruning
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Multistage processes—manufacturing lines, multi-step services, layered systems—generate data whose variables already carry a known sequential hierarchy: later stages cannot cause earlier ones. Standard causal discovery algorithms ignore that hierarchy, so they often invent edges that run backwards through the process and become expensive as the number of stages grows. This paper introduces OCDM, an order-based method that first recovers a causal ordering stage by stage (starting from the last stage and working forward) and then prunes the resulting dense “order-connected” graph with stochastic-gated neural networks. On both synthetic layered graphs and a near-real manufacturing simulator, the stage-aware order plus the neural pruning recovers the true causal graph more faithfully, and usually faster, than existing constraint-based, score-based and order-based baselines. The practical payoff is a causal map that respects the known process layout and can therefore be trusted for root-cause analysis or intervention planning.

Core claim

When every variable is known to belong to a stage of a layered process, a score-matching order search that is forced to identify leaves only inside the current stage, followed by stochastic-gated neural-network pruning of the resulting order-connected graph, recovers the underlying directed acyclic graph more accurately than methods that treat the variables as an unstructured bag.

What carries the argument

Stage-wise leaf identification inside a DiffAN-style score-matching order search, producing an order-connected graph that is then sparsified by stochastic-gated neural networks (STG-NN) whose continuous Bernoulli gates perform L0-style parent selection.

Load-bearing premise

The data must come from a nonlinear additive Gaussian noise model whose variables already sit in known stages of a layered DAG; if either the additive-noise form or the stage labels are badly wrong, both the leaf test and the stage-wise search break.

What would settle it

Generate layered multistage data from a non-additive or non-Gaussian noise model (or deliberately scramble a large fraction of stage labels) and check whether OCDM’s topological-order divergence and structural Hamming distance still beat the best uninformed baselines; if they do not, the claimed advantage disappears.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

2 major / 4 minor

Summary. The paper proposes OCDM, an order-based causal discovery method for multistage processes whose data-generating process is assumed to be a nonlinear additive Gaussian noise model that forms a layered DAG (Sec. III-A, Eq. 1). Stage membership of every variable is treated as known process knowledge. OCDM first runs a stage-wise leaf-node search (Algorithm 1, lines 1–14) that restricts DiffAN-style score-matching order search to the current stage, thereby guaranteeing that the recovered topological order respects the known stage hierarchy. The resulting order-connected graph (Def. 1) is then pruned by training, for each node, an STG-NN whose stochastic gates (Eq. 4) perform L0-relaxed variable selection; edges whose final gate collapses to zero are discarded (Algorithm 1, lines 17–20). Experiments on synthetic layered Erdős–Rényi graphs (few-stage and many-stage), the same graphs with noisy stage labels or mixed continuous/binary variables, and the expert-defined causalAssembly manufacturing simulator show that OCDM attains lower topological-order divergence and better AUROC/AUPRC/SHD/SID than PC, NOTEARS-MLP, DAG-GNN, GraN-DAG, CAM, SCORE, DAS and vanilla DiffAN.

Significance. Multistage manufacturing, service and information systems are ubiquitous, yet standard causal-discovery algorithms ignore the known stage order and scale poorly. By injecting stage membership into an order-based search and replacing CAM-style spline pruning with a more flexible neural gate, OCDM supplies a practical, computationally competitive pipeline that demonstrably improves recovery of layered DAGs on both controlled synthetic graphs and a realistic industrial simulator. The empirical gains (especially the large reduction in D_top and the favorable SID on causalAssembly) are concrete and reproducible; if the method’s assumptions hold, the work offers a useful tool for root-cause analysis and process optimization in multistage settings.

major comments (2)
  1. Sec. III-D and Algorithm 1 (lines 17–20) replace CAM’s group-LASSO spline selection with STG-NN pruning (Eq. 4). No consistency or parent-set recovery result is supplied under the nonlinear additive model of Eq. 1. The authors themselves note in Sec. V that both STG-NN and CAM “lack theoretical guarantees for recovering the true parent set” and that over-pruning of true edges remains a problem. Consequently the multi-metric gains reported in Figs. 3–5 could be driven by aggressive sparsity rather than faithful recovery of the layered DAG; a modest recovery guarantee, or at least a controlled ablation that quantifies true-positive versus false-positive edge retention as candidate-set size grows, is needed to substantiate the central claim that OCDM recovers the causal structure.
  2. The leaf-identification argument (Var[∂s_i/∂x_i]=0) and the stage-wise search both rest on the additive Gaussian noise model (Eq. 1) and on perfectly known stage labels (Sec. III-A). While the noisy-label and mixed-type experiments (Figs. 6–7) are welcome, they remain limited: noise is injected only by random reassignment of a fraction of labels, and the mixed-type data still use a logistic transform of the same Gaussian-process functions. A clearer statement of the breakdown regimes—e.g., when stage information is only partially known or when non-additive interactions dominate—would strengthen the paper’s applicability claims.
minor comments (4)
  1. Fig. 1 caption refers to “Figures 2 and 3 of [36]”; the adapted figure itself is hard to read at the printed scale—consider enlarging node labels or providing a higher-resolution version.
  2. The computational-complexity claim for STG-NN pruning (O(n·E·L·m²) vs. O(nk²+k³) for CAM) is stated without reporting wall-clock numbers for the low-dimensional Simulation 1; adding those numbers would make the efficiency discussion more complete.
  3. Typographical inconsistencies appear in the arXiv identifier (2607.03971) and in the repeated phrase “to address theses issues” (p. 1); a careful proof-reading pass is warranted.
  4. Hyper-parameter settings for the diffusion model and the STG regularization schedule are deferred to the supplement; a short summary table in the main text would improve reproducibility.

Circularity Check

0 steps flagged

No circularity: OCDM is an algorithmic procedure whose order-search and pruning steps are evaluated against external ground-truth DAGs, not quantities defined by the method itself.

full rationale

The paper's derivation chain is the construction of an order-based algorithm (stage-wise leaf identification via DiffAN score Jacobians under the additive-noise model of Eq. 1, followed by STG-NN pruning of the order-connected graph). Stage membership is supplied as known process knowledge (Sec. III-A), not fitted from the same observations used for scoring. The leaf-identification argument (Var[∂s_i/∂x_i]=0 for leaves) is the standard external result of Rolland et al. (SCORE) and Sanchez et al. (DiffAN); STG-NN gates (Eq. 4) are taken from Yamada et al. All reported metrics (D_top, AUROC, AUPRC, SHD, SID) are computed against independently generated synthetic layered DAGs or the expert-defined causalAssembly graph. No equation equates a claimed prediction or uniqueness result to a fitted input, and no load-bearing uniqueness theorem is imported from the authors' own prior work. The acknowledged absence of parent-set recovery guarantees for the pruning step (Sec. V) is a correctness limitation, not circularity. The derivation is therefore self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

2 free parameters · 4 axioms · 0 invented entities

The central claim rests on three modeling axioms taken from the causal-discovery literature plus the domain fact that stage membership is known. No free parameters are fitted to produce the reported performance numbers; the only free choices are standard neural-network hyper-parameters that affect optimization but not the definition of the target graph. No new physical entities are postulated.

free parameters (2)
  • STG noise variance σ_z² and regularization schedule
    Chosen by the authors for the stochastic gates; values are not derived from first principles and affect which edges survive pruning.
  • Diffusion-model architecture and training epochs for score estimation
    Inherited from DiffAN and tuned for the experiments; different choices can change the estimated Jacobian variances used for leaf identification.
axioms (4)
  • domain assumption Data follow a nonlinear additive Gaussian noise model Xi = fi(Pa(Xi)) + Ui (Eq. 1)
    Required for the score-function leaf-identification argument (Sec. III-C) that underpins the entire order search.
  • domain assumption The true graph is a layered DAG whose stage partition is known a priori
    Stated in Sec. III-A; the stage-wise search algorithm is defined only under this structural prior.
  • domain assumption Causal sufficiency (no latent confounders)
    Explicitly assumed in Sec. III-A; score-matching leaf detection fails under latent confounding.
  • standard math Score matching via diffusion models yields a consistent estimator of the score Jacobian variances
    Taken from DiffAN / SCORE literature and used without re-proof.

reviewed 2026-07-11 · how reviews work

0 comments
Cite this review

Pith. "Pith review of Order-based Causal Discovery for Multistage Processes." pith.science (2026). https://pith.science/paper/DWTUV2K3

@misc{pith2026260703971,
  author       = {Pith},
  title        = {Pith review of: Order-based Causal Discovery for Multistage Processes},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DWTUV2K3}},
  note         = {Machine review of arXiv:2607.03971}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Causality has become an increasingly important tool for gaining a deeper understanding of complex systems. Among various causal analysis methods, causal discovery, which identifies causal relationships among variables from data, has been widely used to uncover underlying causality in diverse processes. However, while multistage processes are prevalent in many fields, existing causal discovery methods may produce counterintuitive results, given the known process knowledge, and may not be computationally efficient for handling large datasets typical of multistage processes. To address this gap, we propose a novel causal discovery method called Order-based Causal Discovery for Multistage Processes (OCDM). OCDM is designed to infer the causal structure of multistage data while preserving their inherent hierarchical and sequential structure by explicitly incorporating process knowledge into the causal discovery process. Specifically, we propose a structural knowledge-informed order-inferring algorithm that infers the causal order of variables by incorporating information about the stage from which each variable originates, based on an order-based causal discovery framework naturally suited for inherently ordered multistage data. Furthermore, to eliminate spurious edges from the initial causal graph generated based on the inferred causal order, we introduce a novel pruning technique using stochastic gated neural networks, which offers greater computational efficiency compared to existing methods. Through experiments on various datasets, we demonstrate that OCDM effectively infers the causal structure of multistage processes, outperforming existing methods.

Figures

Figures reproduced from arXiv: 2607.03971 by Eun-Yeol Ma, Heeyoung Kim, Junsub Jung.

Figure 1
Figure 1. Figure 1: Process line of causalAssembly and its ground-truth DAG (figure adapted from Figures 2 and 3 of [36]). Each stage consists of nodes from two distinct adjacent processes. The size of each node represents the number of its out-edges, while the brightness indicates the number of its in-edges. |Li|: number of nodes in stage Li. view of the causalAssembly setup. The causal relationships among the variables with… view at source ↗
Figure 2
Figure 2. Figure 2: Topological order divergence (Dtop) on order-connected graphs. Lower values are preferred, indicating a more accurate causal order. order-connected graphs with the least number of edges that do not exist in the ground-truth DAG, resulting in the most ac￾curate causal structure prior to pruning. Although all methods performed decently on the low-dimensional dataset (Simula￾tion 1) with a low number of false… view at source ↗
Figure 3
Figure 3. Figure 3: Causal discovery results for Simulation 1 ( [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Causal discovery results for Simulation 2 ( [PITH_FULL_IMAGE:figures/full_fig_p009_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Causal discovery results on the pseudo-real data ( [PITH_FULL_IMAGE:figures/full_fig_p010_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Causal discovery results for Simulation 1 ( [PITH_FULL_IMAGE:figures/full_fig_p011_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Causal discovery results for mixed-type data with [PITH_FULL_IMAGE:figures/full_fig_p012_7.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

39 extracted references · 1 linked inside Pith

  1. [1]

    Quality control and improvement for multistage systems: A survey,

    J. Shi and S. Zhou, “Quality control and improvement for multistage systems: A survey,”IIE transactions, vol. 41, no. 9, pp. 744–753, 2009

    2009

  2. [2]

    Peters, D

    J. Peters, D. Janzing, and B. Schölkopf,Elements of causal inference: foundations and learning algorithms. The MIT Press, 2017

    2017

  3. [3]

    Simultaneous treatment effect estimation and variable selection for observational data,

    E.-Y . Ma, U. Lee, and H. Kim, “Simultaneous treatment effect estimation and variable selection for observational data,”IISE Transactions, vol. Early Access, pp. 1–13, 2024

    2024

  4. [4]

    D’ya like dags? a survey on structure learning and causal discovery,

    M. J. V owels, N. C. Camgoz, and R. Bowden, “D’ya like dags? a survey on structure learning and causal discovery,”ACM Computing Surveys, vol. 55, no. 4, pp. 1–36, 2022

    2022

  5. [5]

    Pearl,Causality

    J. Pearl,Causality. New York, NY: Cambridge Univer- sity Press, 2009

    2009

  6. [6]

    Causal structure-based root cause analysis of outliers,

    K. Budhathoki, L. Minorics, P. Blöbaum, and D. Janzing, “Causal structure-based root cause analysis of outliers,” inProceedings of the 39th International Conference on Machine Learning. Baltimore, Maryland, USA: PMLR, 2022, pp. 2357–2369

    2022

  7. [7]

    Contextual anomaly detection for high-dimensional data using dirichlet process variational IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING 13 autoencoder,

    H. Kim and H. Kim, “Contextual anomaly detection for high-dimensional data using dirichlet process variational IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING 13 autoencoder,”IISE Transactions, vol. 55, no. 5, pp. 433– 444, 2023

    2023

  8. [8]

    Application of ker- nel principal component analysis to multi-characteristic parameter design problems,

    W. Soh, H. Kim, and B.-J. Yum, “Application of ker- nel principal component analysis to multi-characteristic parameter design problems,”Annals of Operations re- search, vol. 263, no. 1, pp. 69–91, 2018

    2018

  9. [9]

    Fault detection based on causal discovery and graph convolutional network for complex mechatronic systems,

    C. Wang and J. Liu, “Fault detection based on causal discovery and graph convolutional network for complex mechatronic systems,”IEEE Transactions on Reliability, vol. Early Access, pp. 1–12, 2024

    2024

  10. [10]

    Prediction of highly imbalanced semiconductor chip-level defects in module tests using multimodal fusion and logit adjustment,

    H. Cho, W. Koo, and H. Kim, “Prediction of highly imbalanced semiconductor chip-level defects in module tests using multimodal fusion and logit adjustment,” IEEE Transactions on Semiconductor Manufacturing, vol. 36, no. 3, pp. 425–433, 2023

    2023

  11. [11]

    Why did the distribution change?

    K. Budhathoki, D. Janzing, P. Bloebaum, and H. Ng, “Why did the distribution change?” inProceedings of the 24th International Conference on Artificial Intelligence and Statistics (AISTATS). San Diego, California, USA: PMLR, 2021, pp. 1666–1674

    2021

  12. [12]

    Ordering-based search: a simple and effective algorithm for learning bayesian networks,

    M. Teyssier and D. Koller, “Ordering-based search: a simple and effective algorithm for learning bayesian networks,” inProceedings of the Twenty-First Confer- ence on Uncertainty in Artificial Intelligence, Edinburgh, Scotland, 2005, pp. 584–590

    2005

  13. [13]

    Score matching enables causal discovery of nonlinear additive noise models,

    P. Rolland, V . Cevher, M. Kleindessner, C. Russell, D. Janzing, B. Schölkopf, and F. Locatello, “Score matching enables causal discovery of nonlinear additive noise models,” inProceedings of the 39th International Conference on Machine Learning, vol. 162. Baltimore, Maryland, USA: PMLR, 2022, pp. 18 741–18 753

    2022

  14. [14]

    Diffusion models for causal discovery via topological ordering,

    P. Sanchez, X. Liu, A. Q. O’Neil, and S. A. Tsaftaris, “Diffusion models for causal discovery via topological ordering,” inThe Eleventh International Conference on Learning Representations, Kigali, Rwanda, 2023

    2023

  15. [15]

    Cam: Causal addi- tive models, high-dimensional order search and penalized regression,

    P. Bühlmann, J. Peters, and J. Ernest, “Cam: Causal addi- tive models, high-dimensional order search and penalized regression,”The Annals of Statistics, vol. 42, no. 6, pp. 2526–2556, 2014

    2014

  16. [16]

    Feature selection using stochastic gates,

    Y . Yamada, O. Lindenbaum, S. Negahban, and Y . Kluger, “Feature selection using stochastic gates,” inProceedings of the 37th International Conference on Machine Learn- ing. Online: PMLR, 2020, pp. 10 648–10 659

    2020

  17. [17]

    An algorithm for fast recov- ery of sparse causal graphs,

    P. Spirtes and C. Glymour, “An algorithm for fast recov- ery of sparse causal graphs,”Social Science Computer Review, vol. 9, no. 1, pp. 62–72, 1991

    1991

  18. [18]

    On nonparametric conditional inde- pendence tests for continuous variables,

    C. Li and X. Fan, “On nonparametric conditional inde- pendence tests for continuous variables,”Wiley Interdisci- plinary Reviews: Computational Statistics, vol. 12, no. 3, p. e1489, 2020

    2020

  19. [19]

    Optimal structure identification with greedy search,

    D. M. Chickering, “Optimal structure identification with greedy search,”Journal of Machine Learning Research, vol. 3, no. Nov, pp. 507–554, 2002

    2002

  20. [20]

    Dags with no tears: Continuous optimization for structure learning,

    X. Zheng, B. Aragam, P. K. Ravikumar, and E. P. Xing, “Dags with no tears: Continuous optimization for structure learning,”Advances in neural information processing systems, vol. 31, pp. 9492–9503, 2018

    2018

  21. [21]

    Gradient-based neural dag learning,

    S. Lachapelle, P. Brouillard, T. Deleu, and S. Lacoste- Julien, “Gradient-based neural dag learning,” inThe Eighth International Conference on Learning Represen- tations, Virtual Conference, 2019

    2019

  22. [22]

    Dag-gnn: Dag struc- ture learning with graph neural networks,

    Y . Yu, J. Chen, T. Gao, and M. Yu, “Dag-gnn: Dag struc- ture learning with graph neural networks,” inProceedings of the 36th International Conference on Machine Learn- ing, vol. 97. Long Beach, CA, USA: PMLR, 2019, pp. 7154–7163

    2019

  23. [23]

    Learning sparse nonparametric dags,

    X. Zheng, C. Dan, B. Aragam, P. Ravikumar, and E. Xing, “Learning sparse nonparametric dags,” inPro- ceedings of the 23rd International Conference on Ar- tificial Intelligence and Statistics (AISTATS), vol. 108. Palermo, Italy: PMLR, 2020, pp. 3414–3425

    2020

  24. [24]

    Scalable causal discovery with score matching,

    F. Montagna, N. Noceti, L. Rosasco, K. Zhang, and F. Locatello, “Scalable causal discovery with score matching,” in2nd Conference on Causal Learning and Reasoning, vol. 213. PMLR, 2023, pp. 752–771

    2023

  25. [25]

    Sparse additive models,

    P. Ravikumar, J. Lafferty, H. Liu, and L. Wasserman, “Sparse additive models,”Journal of the Royal Statistical Society Series B: Statistical Methodology, vol. 71, no. 5, pp. 1009–1030, 2009

    2009

  26. [26]

    Causal reasoning meets visual representation learning: A prospective study,

    Y . Liu, Y .-S. Wei, H. Yan, G.-B. Li, and L. Lin, “Causal reasoning meets visual representation learning: A prospective study,”Machine Intelligence Research, vol. 19, no. 6, pp. 485–511, 2022

    2022

  27. [27]

    Cross-modal causal relational reasoning for event-level visual question answering,

    Y . Liu, G. Li, and L. Lin, “Cross-modal causal relational reasoning for event-level visual question answering,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 45, no. 10, pp. 11 624–11 641, 2023

    2023

  28. [28]

    Visual causal scene refinement for video question answering,

    Y . Wei, Y . Liu, H. Yan, G. Li, and L. Lin, “Visual causal scene refinement for video question answering,” in Proceedings of the 31st ACM International Conference on Multimedia, 2023, pp. 377–386

    2023

  29. [29]

    Cross-modal causal relation alignment for video ques- tion grounding,

    W. Chen, Y . Liu, B. Chen, J. Su, Y . Zheng, and L. Lin, “Cross-modal causal relation alignment for video ques- tion grounding,” inProceedings of the Computer Vision and Pattern Recognition Conference, 2025, pp. 24 087– 24 096

    2025

  30. [30]

    Cross-modal causal representation learning for radiology report generation,

    W. Chen, Y . Liu, C. Wang, J. Zhu, G. Li, C.-L. Liu, and L. Lin, “Cross-modal causal representation learning for radiology report generation,”IEEE Transactions on Image Processing, 2025

    2025

  31. [31]

    Masked gradient-based causal structure learning,

    I. Ng, S. Zhu, Z. Fang, H. Li, Z. Chen, and J. Wang, “Masked gradient-based causal structure learning,” in Proceedings of the 2022 SIAM International Conference on Data Mining (SDM). SIAM, 2022, pp. 424–432

    2022

  32. [32]

    Nonlinear causal discovery with additive noise models,

    P. Hoyer, D. Janzing, J. M. Mooij, J. Peters, and B. Schölkopf, “Nonlinear causal discovery with additive noise models,”Advances in neural information process- ing systems, vol. 21, pp. 689–696, 2008

    2008

  33. [33]

    Causal discovery with continuous additive noise mod- els,

    J. Peters, J. M. Mooij, D. Janzing, and B. Schölkopf, “Causal discovery with continuous additive noise mod- els,”Journal of Machine Learning Research, vol. 15, no. 58, pp. 2009–2053, 2014

    2009

  34. [34]

    Hierachical drawing algo- rithms,

    P. Healy and N. S. Nikolov, “Hierachical drawing algo- rithms,” inHandbook of graph drawing and visualiza- tion, R. Tamassia, Ed. CRC press, 2013, ch. 13, pp. 409–453. IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING 14

    2013

  35. [35]

    Regression shrinkage and selection via the lasso,

    R. Tibshirani, “Regression shrinkage and selection via the lasso,”Journal of the Royal Statistical Society Series B: Statistical Methodology, vol. 58, no. 1, pp. 267–288, 1996

    1996

  36. [36]

    causalassembly: Generating realistic production data for benchmarking causal dis- covery,

    K. Göbler, T. Windisch, M. Drton, T. Pychynski, M. Roth, and S. Sonntag, “causalassembly: Generating realistic production data for benchmarking causal dis- covery,” in3rd Conference on Causal Learning and Reasoning, vol. 236. PMLR, 2024, pp. 609–642

    2024

  37. [37]

    On random graphs,

    P. Erd ˝os and A. Rényi, “On random graphs,”Publica- tiones Mathematicae, vol. 6, pp. 290–297, 1959

    1959

  38. [38]

    Structural intervention dis- tance for evaluating causal graphs,

    J. Peters and P. Bühlmann, “Structural intervention dis- tance for evaluating causal graphs,”Neural Computation, vol. 27, no. 3, pp. 771–799, 2015

    2015

  39. [39]

    Score matching through the roof: linear, nonlinear, and latent variables causal discovery,

    F. Montagna, P. M. Faller, P. Blöbaum, E. Kirschbaum, and F. Locatello, “Score matching through the roof: linear, nonlinear, and latent variables causal discovery,” arXiv preprint arXiv:2407.18755, 2024

    Pith/arXiv arXiv 2024

This paper was first reviewed by grok-4.5 on July 11, 2026.