Pith. sign in

REVIEW 3 major objections 4 minor 51 references

Inexact Column Generation for Bayesian Network Structure Learning via Difference-of-Submodular Optimization

T0 review · 3 major / 4 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read Column generation for Bayesian networks can be made practical by solving its pricing problem as a difference of submodular functions with DCA.

desk verdict The DS reformulation of the BNSL pricing problem is sound, but the DCA subgradient formulas in Section 3.2 are missing the N/2 and N scaling factors, so the algorithm as written does not minimize the stated objective. read the letter →

arxiv 2505.11089 v2 pith:RSARMU7I submitted 2025-05-16 stat.ML cs.LGmath.OC

Yiran Yang , Rui Chen This is my paper · ORCID
classification stat.MLcs.LGmath.OC MSC 90C2790C1190C25
keywords Bayesiannetworkstructurelearningcolumngenerationdifference-of-submodularoptimizationDCAl0-penalizedlikelihoodintegerprogrammingrestrictedmasterheuristicLovászextension
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper claims that the pricing step of column generation for score-based Bayesian network structure learning, usually the computational bottleneck, can be rewritten exactly as optimizing a difference of two submodular set functions for ℓ0-penalized Gaussian and multinomial likelihood scores. Because submodular functions have convex Lovász extensions, the difference can be attacked with the Difference of Convex Algorithm, which runs fast and yields candidate parent sets with negative reduced cost. The authors combine this inexact pricing with row and column generation and a restricted master IP heuristic into CG-DCA. On continuous Gaussian data, CG-DCA finds higher BIC scores than the exact IP solver GOBNILP, especially as graph density grows, and matches constraint-based and hybrid baselines in recovery quality. If right, the value is a substantially cheaper pricing step for an NP-hard search problem, making score-based structure learning viable on larger and denser instances.

What carries the argument

The load-bearing object is the Lovász extension of a submodular set function: a convex function on the unit cube that agrees with the set function at 0/1 points. This extension turns the difference-of-submodular pricing objective $\min_J [g(J)-f(J)]$ into a difference-of-convex program $\min_x g^L(x)-f^L(x)$, which DCA solves by linearizing $f^L$ at the current point and minimizing the resulting convex approximation. Because a subgradient of the Lovász extension of $f$ can be evaluated through function evaluations of $f$, each DCA iteration is cheap; for Gaussian scores the paper computes all needed subgradients from a single Cholesky factorization of the permuted covariance matrix, which is what makes the pricing loop fast enough to drive column generation.

What would settle it

Take one Gaussian instance where CG-DCA terminates quickly (say n = 25, d = 2, N = 5000) and rerun the identical row-and-column generation with DCA pricing replaced by an exact pricing oracle that enumerates all parent sets up to size 4; if the oracle finds a negative reduced cost after CG-DCA has stopped, or the final restricted master IP has a strictly higher BIC score, then the claim that DCA pricing is sufficient for high-quality Gaussian BNSL fails on that instance.

Watch

Extended reading notes

Core claim

The central claim is that, for ℓ0-penalized likelihood scores, the reduced cost of a candidate parent set $J$ for node $i$ in the column-generation pricing problem, $z_i(J;\lambda^*)=-\mathrm{score}_i(J)+\lambda^*_i+\sum_{C\in\hat{C}: i\in C, J\cap C\ne \emptyset}\lambda^*_C$, can be written exactly as $g(J)-f(J)$ with both $g$ and $f$ submodular. For Gaussian data the proof uses the Schur-complement identity $\hat\sigma^2_{i\leftarrow J}=\det(\hat\Sigma_{J\cup\{i\},J\cup\{i\}})/\det(\hat\Sigma_{J,J})$ together with the fact that $\log\det$ of a principal submatrix is submodular; for multinomial data it uses entropy submodularity and treats the parameter-count penalty $-\Lambda(a_i-1)\prod_{j\in J}a_j$ as part of the subtracted supermodular term. Since the Lovász extension converts each submodular function into a convex function, minimizing $g-f$ becomes a difference-of-convex program, and the Difference of Convex Algorithm (DCA) produces a non-increasing sequence of objective values and, at integer points, recovers the original set-function value. The paper's experiments show that this inexact pricing is fast and accurate enough on Gaussian data that the overall CG-DCA method beats the exact GOBNILP solver in BIC score within a 3-hour budget and matches constraint-based and hybrid methods in graph recovery quality.

Load-bearing premise

The load-bearing premise is empirical: DCA's local solutions to the difference-of-submodular pricing problem must have negative reduced costs often enough for the restricted master heuristic to find good DAGs, since DCA only guarantees that its objective sequence does not increase and gives no global optimality certificate.

Editorial extensions

If this is right

  • On simulated Gaussian data with n = 15, 20, and 25, CG-DCA achieves BIC scores at or near the true-graph score and smaller average BIC gaps than GOBNILP, with the advantage growing as the average in-degree increases.
  • Pricing with DCA is orders of magnitude faster than the MINLP pricing baseline: it recovers the Gaussian.test ground truth in 1.2 seconds versus 19.2 seconds, and per-node pricing drops from about 35.1 seconds to 0.3 seconds on an n = 20, d = 1 instance.
  • Collecting intermediate DCA solutions as candidate columns and using warm-started restricted master IPs yields valid DAGs, so the approach works as a primal heuristic with acyclicity enforced through lazy cluster constraints.
  • On discrete data the supermodular penalty term $(a_i-1)\prod_{j\in J}a_j$ in the multinomial score makes DCA less effective, and GOBNILP outperforms CG-DCA on the LUCAS, INSURANCE, and ALARM benchmarks.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A direct extension the paper leaves implicit is to embed DCA pricing inside exact branch-and-price: because column generation only needs exact pricing near termination, a DCA-generated negative-reduced-cost column could be accepted as a warm candidate and certified later by an exact oracle, potentially retaining the speedup while restoring optimality guarantees.
  • The same $g-f$ decomposition should transfer to other ℓ0-penalized scores whose log-likelihood term is submodular in the parent set, such as some exponential-family or nonparametric scores; the two propositions in the paper are templates, not just special cases.
  • The reported precision/recall pattern—high recall, moderate precision, denser recovered graphs on Gaussian data—suggests that inexact pricing biases column generation toward adding many parent sets; a testable prediction is that the number of negative-reduced-cost columns added per iteration is larger under DCA pricing than under exact pricing.
  • The switch from random to warm-start initialization after about 50 candidate parent sets is an algorithmic choice that likely carries much of the empirical speedup; varying that threshold and measuring BIC and runtime would isolate how much of the gain comes from the DS reformulation versus the initialization schedule.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. The paper proposes an inexact column generation (CG) method for Bayesian network structure learning (BNSL). For l0-penalized Gaussian and multinomial likelihood scores, it rewrites the CG pricing problem as a difference of submodular (DS) functions (Propositions 1 and 2) and applies the Difference of Convex Algorithm (DCA) to generate columns with negative reduced costs. The resulting CG-DCA is compared against GOBNILP, CG-MINLP, hill climbing, stable-PC, and MMHC on simulated Gaussian data (n=15,20,25; average in-degree 0.5-2) and on three small discrete benchmarks. The authors report better BIC scores than GOBNILP on Gaussian instances within a 3-hour time limit, and structural accuracy competitive with constraint-based and hybrid baselines; they also report that DCA is ineffective for discrete scores.

Significance. The DS reformulation of the pricing problem is a clean and checkable contribution: it replaces an MINLP pricing subproblem with submodular optimization, for which efficient local methods such as DCA exist, and the algebraic derivations in Propositions 1-2 are transparent and use standard submodularity arguments. If the implementation matched the stated objective, the approach would be a useful heuristic for Gaussian BNSL and a step toward more scalable exact branch-price-and-cut. However, the paper does not release code, the DCA subgradient formulas in Section 3.2 contain a scaling error relative to the stated DS decomposition, and the empirical claims are weakened by the absence of error bars and by the 3-hour time limits that often terminate GOBNILP without proof of optimality. The discrete results are explicitly acknowledged by the authors to degrade, so the demonstrated practical contribution is essentially limited to Gaussian scores.

major comments (3)
  1. [§3.2] The displayed subgradient formulas for f_L omit the multiplicative constants present in Propositions 1 and 2. In the continuous case, Proposition 1 defines f(J) = (N/2) log det(Σ_{J,J}); hence a subgradient of f_L at x^t along the sorted order has k-th component (N/2)(log det(Σ_{S_k}) − log det(Σ_{S_{k−1}})), but the paper prints log det(Σ_{S_k}) − log det(Σ_{S_{k−1}}) with no factor N/2. In the discrete case, Proposition 2 defines f(J) = N·H(J) − Λ(a_i−1)∏_{j∈J} a_j; the printed subgradient omits N on the entropy differences. Consequently, the vector y^t used in Algorithm 1 Line 5 is not a subgradient of the stated f_L, the subproblem solved is not the DCA linearization of z_L = g_L − f_L, and the monotone non-increase guarantee (4)—which the paper relies on for the convergence of the pricing heuristic—does not follow for the objective of Propositions 1-2. Because no code is released, the reported CG-DCA results cannot be checked to determine whether the experiments used the correct factors or the printed unscaled formulas. This must be corrected and the experiments either re-run with the correct subgradients or explicitly documented as using a differently scaled objective.
  2. [Table 1 / Abstract] The headline claim that CG-DCA 'consistently outperforms GOBNILP in terms of average scores' is a comparison under a 3-hour time limit, not a claim about optimality. Several GOBNILP rows report runtimes at or near 10800 seconds (e.g., the (25,5000,2.0) and (25,20000,2.0) rows show max times of 10801.63s and 10801.71s), so for those instances GOBNILP was terminated without proving optimality and the reported BIC gaps are upper bounds on the achievable gaps. The abstract and Section 5.2 should qualify the claim as 'within the time budget.' In addition, Table 1 reports (min, max) and average over ten instances without standard deviations or any measures of variability; with only ten replications and gaps as small as 0.00–0.03 in several rows, the phrase 'consistently outperforms' is not statistically supported. Please add per-cell variation (e.g., standard errors), report how many instances hit the time limit for each solver, and adjust the wording accordingly.
  3. [§5.3, §6] The paper's own Section 5.3 states that for the discrete BIC score the 'highly supermodular penalization term ... makes DCA less effective,' and Table 3 shows CG-DCA with SHD 35–37 on ALARM versus GOBNILP's 7 and precision 0.46 versus 0.87. The conclusion acknowledges limited scalability for discrete data. Given this, the abstract's phrase 'for the general class of ℓ0-penalized likelihood scores' overstates the demonstrated scope: the DS reformulation in Propositions 1-2 is general, but the proposed algorithm is shown to be effective only for Gaussian scores, while the discrete experiments serve as a negative result for the DCA approach. The abstract and introduction should temper the general-claim language, or the discrete results should be framed explicitly as a limitation rather than as evidence of viability.
minor comments (4)
  1. [Equation (3)] The quantity z_i(J;λ*) in (3) is the negative of the standard reduced cost for a maximization LP: the standard reduced cost of x_{i←J} would be score_i(J) − λ*_i − ∑_{C} λ*_C. Please state this sign convention explicitly to avoid confusion when comparing with the column generation literature.
  2. [Algorithm 1 / Algorithm 2] Algorithm 1 returns a vector x^t, while Algorithm 2 Line 10 uses a parent set 'pa'; the paper does not specify how the binary parent set is recovered from the possibly fractional DCA iterate. Please add a sentence describing the rounding or extreme-point recovery procedure used in the implementation.
  3. [§4.1] The hybrid initialization switches from random to warm-start once the candidate set contains 50 parent sets, but Appendix D.1 only compares the three strategies at this fixed threshold and provides no sensitivity analysis. Please report results for a range of threshold values or justify the chosen value.
  4. [§5.1] The word 'inaccurrate' in the sentence about MINLP numerical issues should be spelled 'inaccurate'.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: the DS pricing reformulation is a direct algebraic identity and the empirical claims are benchmarked against external methods.

full rationale

The central reduction (Propositions 1 and 2) rewrites the pricing objective (3) by substituting the closed-form Gaussian conditional-variance expression log(sigma^2_{i<-J}) = logdet(Sigma_{J union i,J union i}) - logdet(Sigma_{J,J}) and the multinomial identity sum # log(#) = N(H(J) - H(J union {i})), then separates the resulting expression into g(J) - f(J). No term is fitted to the downstream BIC scores, no parameter is estimated from the benchmark target, and no prediction is a renamed input. The submodularity facts are taken from external literature (e.g., Kelmans-Kimelfeld for logdet, Fujishige for entropy), and the CG/pricing framework is attributed to prior external work by Cussens. The only possible self-citation, [12], appears in a general list of IP-based structure-learning formulations and plays no role in the DS derivation or in the experimental advantage. The empirical section compares against GOBNILP, PC, MMHC, HC, and ground-truth graphs, so the reported superiority is an externally evaluated outcome rather than a constructed one. The skeptical observation that the Section 3.2 subgradient formulas omit the N/2 (Gaussian) and N (discrete) scale factors is a potential implementation mismatch between the stated objective and Algorithm 1; it is a correctness concern, not an instance of circularity, because it does not make the DS objective equal to its inputs and does not define any quantity in terms of the target result. Section 5.3 and Section 6 candidly report the discrete-data limitation, further confirming that the claims are not protected by a circular argument.

Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

The method introduces no new entities. It relies on standard submodularity and convexity results and on domain assumptions about Gaussian and multinomial data. The two hand-set hyperparameters are not load-bearing in the theoretical claim but affect reproducibility.

free parameters (2)
  • Hybrid initialization switch threshold = 50
    The number of candidate parent sets after which DCA initialization switches from random to warm-start (Section 4.1); hand-set and not tuned or sensitivity-analyzed.
  • DCA stopping threshold epsilon = not stated
    Controls when Algorithm 1 terminates; the paper does not report the value used in experiments, so results are not exactly reproducible from the text.
assumptions (5)
  • domain assumption For continuous data, empirical covariance matrices are positive definite for every parent set J union {i} considered, so logdet is finite.
    Invoked in Proposition 1 and in the DCA subgradient formulas in Section 3.2. It fails when |J| is at least the sample size or under collinearity; experiments use N much larger than n.
  • domain assumption Gaussian conditional distributions with mean-zero variables are the correct model for continuous nodes.
    Used in Proposition 1 and the local score formula in Appendix A.1; standard for BNSL with continuous data.
  • standard math Entropy and logdet of principal submatrices are submodular, and the Lovasz extension of a submodular function is convex.
    Used in Propositions 1 and 2 and Section 2.3; cited to references [24], [34], and [36].
  • standard math DCA converges monotonically and its subproblem minima can be taken at integer points of the unit box.
    Used to justify solving the pricing problem via Algorithm 1; cited to [22] and [30]. The monotonicity claim is in Eq. (4).
  • standard math Cluster constraints with nonnegative duals lambda_C fully encode acyclicity in the IP formulation.
    Used in formulation (2) and pricing problem (3); standard from [15] and [31].

how reviews work

0 comments
Cite this review

Pith. "Pith review of Inexact Column Generation for Bayesian Network Structure Learning via Difference-of-Submodular Optimization." pith.science (2026). https://pith.science/paper/RSARMU7I

@misc{pith2026250511089,
  author       = {Pith},
  title        = {Pith review of: Inexact Column Generation for Bayesian Network Structure Learning via Difference-of-Submodular Optimization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RSARMU7I}},
  note         = {Machine review of arXiv:2505.11089}
}
abstract

In this paper, we consider a score-based Integer Programming (IP) approach for solving the Bayesian Network Structure Learning (BNSL) problem. State-of-the-art BNSL IP formulations suffer from the exponentially large number of variables and constraints. A standard approach in IP to address such challenges is to employ row and column generation techniques, which dynamically generate rows and columns, while the complex pricing problem remains a computational bottleneck for BNSL. For the general class of $\ell_0$-penalized likelihood scores, we show how the pricing problem can be reformulated as a difference of submodular optimization problem, and how the Difference of Convex Algorithm (DCA) can be applied as an inexact method to efficiently solve the pricing problems. Empirically, we show that, for continuous Gaussian data, our row and column generation approach yields solutions with higher quality than state-of-the-art score-based approaches, especially when the graph density increases, and achieves comparable performance against benchmark constraint-based and hybrid approaches, even when the graph size increases.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

51 extracted references · 47 canonical work pages

  1. [1]

    H. Akaike. A new look at the statistical model identification problem.IEEE Transactions on Automatic Control, 19:716, 1974

    work page 1974

  2. [2]

    L. T. H. An and P. D. Tao. The DC (difference of convex functions) programming and DCA revisited with DC models of real world nonconvex optimization problems.Annals of Operations Research, 133:23–46, 2005

    work page 2005

  3. [3]

    S. A. Andersson, D. Madigan, and M. D. Perlman. A characterization of markov equivalence classes for acyclic digraphs.The Annals of Statistics, 25(2):505–541, 1997

    work page 1997

  4. [4]

    Bagirov, N

    A. Bagirov, N. Karmitsa, and M. M. Mäkelä.Introduction to Nonsmooth Optimization: Theory, Practice and Software, volume 12. Springer, 2014

    work page 2014

  5. [5]

    Barnhart, E

    C. Barnhart, E. L. Johnson, G. L. Nemhauser, M. W. Savelsbergh, and P. H. Vance. Branch-and- price: Column generation for solving huge integer programs.Operations Research, 46(3):316– 329, 1998

    work page 1998

  6. [6]

    Bartlett and J

    M. Bartlett and J. Cussens. Integer linear programming for the bayesian network structure learning problem.Artificial Intelligence, 244:258–271, 2017

    work page 2017

  7. [7]

    I. A. Beinlich, H. J. Suermondt, R. M. Chavez, and G. F. Cooper. The alarm monitoring system: A case study with two probabilistic inference techniques for belief networks. InAIME 89: Second European Conference on Artificial Intelligence in Medicine, London, August 29th–31st

    work page

  8. [8]

    Bertsimas and J

    D. Bertsimas and J. N. Tsitsiklis.Introduction to linear optimization, volume 6. Athena scientific Belmont, MA, 1997

    work page 1997

Show all 51 references
  1. [9]

    Bhattacharya, T

    R. Bhattacharya, T. Nagarajan, D. Malinsky, and I. Shpitser. Differentiable causal discovery under unmeasured confounding. InInternational Conference on Artificial Intelligence and Statistics, pages 2314–2322. PMLR, 2021

    2021

  2. [10]

    Binder, D

    J. Binder, D. Koller, S. Russell, and K. Kanazawa. Adaptive probabilistic networks with hidden variables.Machine Learning, 29:213–244, 1997

    1997

  3. [11]

    W. Buntine. Theory refinement on bayesian networks. InUncertainty in artificial intelligence, pages 52–60. Elsevier, 1991

    1991

  4. [12]

    R. Chen, S. Dash, and T. Gao. Integer programming for causal structure learning in the presence of latent variables. InInternational Conference on Machine Learning, pages 1550–1560. PMLR, 2021

    2021

  5. [13]

    D. M. Chickering, D. Heckerman, and C. Meek. Large-sample learning of bayesian networks is NP-hard.Journal of Machine Learning Research, 5(Oct):1287–1330, 2004

    2004

  6. [14]

    Colombo and M

    D. Colombo and M. H. Maathuis. Order-independent constraint-based causal structure learning. Journal of Machine Learning Research, 15(1):3741–3782, 2014

    2014

  7. [15]

    J. Cussens. Bayesian network learning with cutting planes. InProceedings of the Twenty- Seventh Conference on Uncertainty in Artificial Intelligence, UAI’11, page 153–160, Arlington, Virginia, USA, 2011. AUAI Press

    2011

  8. [16]

    J. Cussens. Column generation for exact BN learning: Work in progress. InProc. ECAI-2012 workshop on COmbining COnstraint solving with MIning and LEarning (CoCoMile 2012), 2012

    2012

  9. [17]

    J. Cussens. GOBNILP: Learning Bayesian network structure with integer programming. In International Conference on Probabilistic Graphical Models, pages 605–608. PMLR, 2020

    2020

  10. [18]

    J. Cussens. Branch-price-and-cut for causal discovery. InConference on Causal Learning and Reasoning, pages 642–661. PMLR, 2023. 11

    2023

  11. [19]

    Desaulniers, J

    G. Desaulniers, J. Desrosiers, and M. M. Solomon.Column Generation, volume 5. Springer Science & Business Media, 2006

    2006

  12. [20]

    Desrochers, J

    M. Desrochers, J. Desrosiers, and M. Solomon. A new optimization algorithm for the vehicle routing problem with time windows.Operations Research, 40(2):342–354, 1992

    1992

  13. [21]

    Drton and M

    M. Drton and M. H. Maathuis. Structure learning in graphical modeling.Annual Review of Statistics and Its Application, 4(1):365–393, 2017

    2017

  14. [22]

    El Halabi, G

    M. El Halabi, G. Orfanides, and T. Hoheisel. Difference of submodular minimization via DC programming. InInternational Conference on Machine Learning, pages 9172–9201. PMLR, 2023

    2023

  15. [23]

    LUCAS (LUng CAncer Simple set)

    ETH Zurich. LUCAS (LUng CAncer Simple set). https://www.causality.inf.ethz.ch/ data/LUCAS.html

  16. [24]

    Fujishige

    S. Fujishige. Polymatroidal dependence structure of a set of random variables.Information and Control, 39(1):55–72, 1978

    1978

  17. [25]

    Fujishige.Submodular Functions and Optimization, volume 58

    S. Fujishige.Submodular Functions and Optimization, volume 58. Elsevier, 2005

    2005

  18. [26]

    P. C. Gilmore and R. E. Gomory. A linear programming approach to the cutting-stock problem. Operations Research, 9(6):849–859, 1961

    1961

  19. [27]

    C. N. Glymour and G. F. Cooper.Computation, Causation, and Discovery. AAAI Press, 1999

    1999

  20. [28]

    Heckerman, D

    D. Heckerman, D. Geiger, and D. M. Chickering. Learning bayesian networks: The combination of knowledge and statistical data.Machine Learning, 20:197–243, 1995

    1995

  21. [29]

    Heinze-Deml, M

    C. Heinze-Deml, M. H. Maathuis, and N. Meinshausen. Causal structure learning.Annual Review of Statistics and Its Application, 5(1):371–391, 2018

    2018

  22. [30]

    Iyer and J

    R. Iyer and J. Bilmes. Algorithms for approximate minimization of the difference between submodular functions, with applications.Uncertainty in Artificial Intelligence (UAI), 2012

    2012

  23. [31]

    Jaakkola, D

    T. Jaakkola, D. Sontag, A. Globerson, and M. Meila. Learning Bayesian network structure using LP relaxations. InProceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics, pages 358–365. JMLR Workshop and Conference Proceedings, 2010

    2010

  24. [32]

    Joncour, S

    C. Joncour, S. Michel, R. Sadykov, D. Sverdlov, and F. Vanderbeck. Column generation based primal heuristics.Electronic Notes in Discrete Mathematics, 36:695–702, 2010

    2010

  25. [33]

    J. E. Kelley, Jr. The cutting-plane method for solving convex programs.Journal of the society for Industrial and Applied Mathematics, 8(4):703–712, 1960

    1960

  26. [34]

    A. K. Kelmans and B. N. Kimelfeld. Multiplicative submodularity of a matrix’s principal minor as a function of the set of its rows and some combinatorial applications.Discrete Mathematics, 44(1):113–116, 1983

    1983

  27. [35]

    Koivisto and K

    M. Koivisto and K. Sood. Exact bayesian structure discovery in bayesian networks.Journal of Machine Learning Research, 5(May):549–573, 2004

    2004

  28. [36]

    L. Lovász. Submodular functions and convexity.Mathematical Programming The State of the Art: Bonn 1982, pages 235–257, 1983

    1982

  29. [37]

    M. E. Lübbecke and J. Desrosiers. Selected topics in column generation.Operations Research, 53(6):1007–1023, 2005

    2005

  30. [38]

    C. Meek. Causal inference and causal explanation with background knowledge.arXiv preprint arXiv:1302.4972, 2013

    2013 arXiv

  31. [39]

    G. L. Nemhauser and L. A. Wolsey. Maximizing submodular set functions: formulations and analysis of algorithms. InNorth-Holland Mathematics Studies, volume 59, pages 279–301. Elsevier, 1981. 12

    1981

  32. [40]

    Pearl.Causality: Models, Reasoning and Inference

    J. Pearl.Causality: Models, Reasoning and Inference. Cambridge University Press, 2000

    2000

  33. [41]

    Potthoff, D

    D. Potthoff, D. Huisman, and G. Desaulniers. Column generation with dynamic duty selection for railway crew rescheduling.Transportation Science, 44(4):493–505, 2010

    2010

  34. [42]

    G. Schwarz. Estimating the dimension of a model.The Annals of Statistics, pages 461–464, 1978

    1978

  35. [43]

    M. Scutari. Learning Bayesian networks with the bnlearn R package.Journal of Statistical Software, 35:1–22, 2010

    2010

  36. [44]

    Silander and P

    T. Silander and P. Myllymaki. A simple approach for finding the globally optimal bayesian network structure.arXiv preprint arXiv:1206.6875, 2012

    2012 arXiv

  37. [45]

    Spirtes, C

    P. Spirtes, C. N. Glymour, and R. Scheines.Causation, Prediction, and Search. MIT Press, 2000

    2000

  38. [46]

    Tillman and P

    R. Tillman and P. Spirtes. Learning equivalence classes of acyclic models with latent and selection variables from multiple datasets with overlapping variables. InProceedings of the Fourteenth International Conference on Artificial Intelligence and Statistics, pages 3–15. JMLR...

    2011

  39. [47]

    Tsamardinos, L

    I. Tsamardinos, L. E. Brown, and C. F. Aliferis. The max-min hill-climbing bayesian network structure learning algorithm.Machine Learning, 65:31–78, 2006

    2006

  40. [48]

    Van de Geer and P

    S. Van de Geer and P. Bühlmann. L0-penalized maximum likelihood for sparse directed acyclic graphs.The Annals of Statistics, 43(2):536–567, 2013

    2013

  41. [49]

    Y . Yu, J. Chen, T. Gao, and M. Yu. DAG-GNN: DAG structure learning with graph neural networks. InInternational Conference on Machine Learning, pages 7154–7163. PMLR, 2019

    2019

  42. [50]

    J. Zhang. On the completeness of orientation rules for causal discovery in the presence of latent confounders and selection bias.Artificial Intelligence, 172(16-17):1873–1896, 2008. A Definitions A.1 The Data Matrix and Scoring Functions We denoteD as the dataset, which is anN...

    2008

  43. [1989]

    Springer, 1989

    Proceedings, pages 247–256. Springer, 1989

    1989

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.