Random Sum-Product Forests with Residual Links

Alejandro Molina; Fabrizio Ventola; Karl Stelzner; Kristian Kersting

REVIEW 2 major objections 6 minor 42 references

Random Sum-Product Forests with Residual Links

T0 review · 2 major / 6 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read Randomly generated sum-product networks, combined into a forest and joined by residual links, match the accuracy of structure-learned density estimators.

desk verdict A genuinely new residual-link mechanism for combining SPNs, with plausible empirical gains, but Algorithm 4 as written can create invalid SPNs and the evaluation lacks error bars. read the letter →

arxiv 1908.03250 v1 pith:OR4Y23OR submitted 2019-08-08 cs.LG cs.AIstat.ML

Fabrizio Ventola , Karl Stelzner , Alejandro Molina , Kristian Kersting This is my paper · ORCID

classification cs.LGcs.AIstat.ML

keywords sum-productnetworksdensityestimationensemblelearningresidualtractableprobabilisticmodelsstructurerandomSPNstructurestestlog-likelihood

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

Scaling sum-product networks has been held back by the cost of learning their structure from data, so this paper proposes to sidestep that cost with randomness plus ensembling. It introduces random sum-product forests (RSPFs), which mix several randomly generated SPNs under a single sum node, and residual sum-product networks (ResSPNs), which add residual links that let one component refine another component's density estimate over the same variables. The reported experiments show that the forest always beats its best individual component, and residual links improve test log-likelihood on five of six datasets, with Jester overfitting. The payoff, if the comparisons hold, is that cheap random components plus residual wiring approach the accuracy of expensive structure learners while keeping exact and tractable inference.

What carries the argument

The load-bearing mechanism is the residual sum-product link: adding a node from another SPN as an extra child of a sum node with the same scope (marginalizing surplus variables when the donor scope is broader), so the receiving node fits a weighted residual $b(z)=a(z)+r(z)$ instead of the full density. This preserves the completeness and decomposability conditions that keep SPNs valid while making the receiving network wider and deeper. The component generator is the ExtraSPN, which replaces LearnSPN's statistical independence tests with random splits controlled by a failure probability $\beta$ and samples the minimum-split-instance threshold $\mu$ to create diversity; the RSPF then mixes the components under one top sum node.

What would settle it

Re-run LearnSPN and ID-SPN on the same six datasets with the same train/validation/test splits and the same evaluation pipeline, then compare average test log-likelihood; if either baseline beats ResSPN on most datasets under identical conditions, the competitiveness claim fails. A cheaper check is to run the RSPF-versus-best-ExtraSPN comparison over many random seeds: if the best ExtraSPN sometimes wins, the claim that RSPF always outperforms its components is not robust.

Watch

Extended reading notes

Core claim

The central claim is that ensembles of randomly structured SPNs repair the false independence assumptions that hurt individual random SPNs. In the paper's experiments, an RSPF of ten ExtraSPNs—randomized, extremely-randomized-tree-style versions of LearnSPN's recursive splitting and clustering—achieves higher average test log-likelihood than the best single ExtraSPN on all six benchmark datasets. Adding residual links yields a ResSPN that improves on the plain RSPF on five of six datasets, with Jester as the overfitting exception. The resulting ResSPN test log-likelihoods are comparable to LearnSPN on several datasets and close to ID-SPN, which the paper reads as evidence that random generation plus ensemble wiring is a practical alternative to expensive structure learning.

Load-bearing premise

The competitiveness claim depends on taking the LearnSPN and ID-SPN test log-likelihood numbers from an earlier publication as directly comparable, without re-running those baselines under the same data splits, preprocessing, and evaluation code, and on reporting single point estimates with no error bars.

Editorial extensions

If this is right

Ensembles of random SPNs can correct wrong independence assumptions made by single components: a trained ResSPN's pairwise mutual information approaches the empirical values on NLTCS.
Residual links increase both the depth and width of the resulting SPN compared with singleton tree-shaped learners, without giving up exact inference.
The framework is general: other structure learners, including the stronger ID-SPN, can be plugged in as components, and different residual-link selection strategies can be substituted.
Because the random splitting procedure has constant rather than quadratic cost, RSPFs and ResSPNs scale naturally to thousands of random variables.
An informed variant that adds a residual link only when it locally improves the training objective is less prone to overfitting and outperforms the RSPF baseline on the tested datasets.

Reading between the lines

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

If residual links serve hierarchical mixture models as they serve neural networks, then iteratively adding links and fine-tuning after each addition should yield a boosting-like gain; the paper tests only one-shot link addition followed by joint EM.
The same-scope-with-marginalization rule suggests a general transfer mechanism between tractable models over partially overlapping variable sets, potentially applicable to other probabilistic circuits beyond SPNs.
Mixing learned and random components in one forest, with residual links connecting them, is a natural untested extension that could combine the strengths of both structure sources.
The Jester overfitting case implies residual links add capacity; a validation-based early stopping rule for link addition might make the improvement uniform, which the reported experiments do not check.

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.

Desk Editor's Note

read the letter

Dear colleague, quick take: the paper has a genuinely new idea—residual links that attach a marginalized copy of a node from another SPN as an extra child—and it shows on standard benchmarks that ensembling random SPNs helps, and residual links help a bit more. The problem is that the algorithm defining residual links has a formal gap that could invalidate the whole model class. I checked Algorithm 4: it iterates over all nodes s1 in BFS(ResSPN), and whenever scope(s1) is a subset of scope(s2), it adds a connection from s1 to a marginalized copy of s2. There is no check that s1 is a sum node. If s1 is a product node, adding a child with scope equal to the product's scope necessarily overlaps with existing children's scopes, violating decomposability. If s1 is a leaf, you're adding a child to a univariate distribution, which is not a valid SPN operation. The prose in Section 4 says "We add Alice's product node as a child to Bob's sum node," so the intended operation is sum-parent only, but the pseudocode omits that restriction. Either the released code follows the pseudocode and the reported likelihoods are not densities, or the code follows the prose and the formal specification is incomplete. Either way, that's load-bearing for a method whose selling point is that the models remain valid, tractable SPNs.

What's good: the residual-link idea is new as far as I know, and the paper is honest about prior ensemble work. RSPF is a simple top-sum ensemble of ExtraSPNs, and Table 2 shows it beats the best ExtraSPN on all six datasets; residual links improve on RSPF on five of six, with Jester as the expected overfitting case. The MI visualization in Figure 3 is a good sanity check, and the code is available.

The other soft spots are secondary but real: no error bars or significance tests anywhere, so the improvements are point estimates; and the LearnSPN/ID-SPN baselines are taken from Rooshenas and Lowd without re-running, so the comparisons could be affected by different preprocessing. None of that kills the story, but it needs addressing.

Who should read it: anyone working on SPN structure learning or ensembles of tractable probabilistic models; the scaling argument is relevant. I'd send it to peer review, but with a request for major revision: correct Algorithm 4, add multiple-seed or bootstrapped error bars, and ideally re-run the baselines. After those fixes, the residual-link contribution is a solid increment to the SPN toolbox.

Referee Report

2 major / 6 minor

Summary. The manuscript introduces Random Sum-Product Forests (RSPFs), an ensemble method that mixes randomly generated ExtraSPNs via a top sum node, and Residual Sum-Product Networks (ResSPNs), which augment one SPN with 'residual links' to nodes of other SPNs, followed by a global mixture. The authors evaluate on six binary density-estimation benchmarks and report that RSPFs outperform the best single ExtraSPN on all datasets, that residual links further improve test likelihoods on five of six datasets (Jester overfits), and that ResSPNs are competitive with LearnSPN and approach ID-SPN. They also propose an informed variant (InfoResSPN) and provide structural statistics indicating that ResSPNs are wider and deeper than tree-shaped learned SPNs.

Significance. If the empirical claims hold, the paper offers a practical and scalable alternative to structure learning for SPNs, and the residual-link idea is a plausible mechanism for sharing context-specific substructures across an ensemble. The approach is simple, the code is released, and the experimental protocol (train/validation/test splits, EM optimization, hyperparameter search) is largely standard. However, the validity of the ResSPN construction is not guaranteed by the algorithm as written, and the comparisons against LearnSPN/ID-SPN rely on externally reported numbers without error bars, so the significance of the stated improvements is currently not fully established.

major comments (2)

[Section 4, Algorithm 4] The pseudocode does not restrict the type of node s1 when adding a residual link. The condition scope(s1) ⊆ scope(s2) can be satisfied by product nodes and leaves. Adding a child to a product node with a scope that overlaps the scopes of existing children violates the decomposability requirement, and adding a child to a leaf is syntactically invalid for a univariate distribution. The surrounding prose ('We add Alice’s product node as a child to Bob’s sum node') suggests the intended operation is limited to sum nodes as parents, but the algorithm as written permits arbitrary node types. Since the ResSPN likelihoods in Tables 2, 4, and 5 depend on the resulting network being a valid, normalized SPN, this gap is load-bearing. The authors should either restrict s1 to sum nodes in Algorithm 4, or provide a proof that adding links to product nodes and leaves preserves decomposability, or re-run the experiments with an implementation that enforces the restriction.
[Section 5.1, Table 2] The LearnSPN and ID-SPN baseline numbers are taken from [Rooshenas and Lowd, 2014] rather than re-computed with the same code, preprocessing, and train/validation/test splits used for the proposed methods. The manuscript does not report standard deviations, confidence intervals, or significance tests for any of the average test log-likelihoods, so it is unclear whether the observed differences (e.g., -6.046 vs. -6.040 for RSPF vs. ResSPN on NLTCS) are meaningful. This undermines the abstract's claim that ResSPNs are 'competitive with commonly used structure learning methods.' The authors should either rerun the baselines in SPFlow, report variance estimates, or soften the comparative claims.

minor comments (6)

[Algorithm 4] The prose says the BFS traversal excludes the root, but the pseudocode in Algorithm 4 does not show this exclusion. Please clarify the discrepancy.
[Section 5.2] The 'best ExtraSPN' is selected using test-set likelihood; this should be based on a validation set to avoid potential bias, even though the direction of the bias appears to favor the ensemble.
[Section 5.4] The statement that InfoResSPN 'achieves overall better accuracy than ResSPN' is not directly supported by a table comparing the two; Table 5 compares InfoResSPN to RSPF. Please add a direct comparison or explicitly note that the claim is based on comparing Table 5 with Table 2.
[Conclusion] The conclusion uses 'RSPNs' where 'RSPFs' is meant ('RSPNs and ResSPNs can be seen as general schemas'). Please correct this typo.
[Table 4] The 'wins' row is not labeled; it appears to count pairwise comparisons between ResSPN and RSPF of the same ensemble size. Please add a caption or explanatory note to make the row self-explanatory.
[Section 4] The term 'residual link' may be misleading because the implementation adds a child to a sum node, which corresponds to a mixture, not an additive residual. The conceptual connection to ResNets is only loose; consider acknowledging this explicitly to avoid overclaiming the analogy.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: RSPF/ResSPN claims are empirical evaluations against external baselines, not derivations that fold target results into inputs.

full rationale

The paper's central claims—RSPFs improve over their ExtraSPN components and residual links further improve RSPFs—are established by held-out test log-likelihood comparisons (Tables 2, 4, and 5) against external baselines, including LearnSPN and ID-SPN as reported by Rooshenas and Lowd (2014). The construction algorithms (Algs. 2–4) take a set of input SPNs and combine, mix, or add residual links to them; no parameter is fitted to a subset of data and then reported as a prediction of that same subset. The only tuning choices, such as the residual-link ratio k in [0.1, 0.2] and the number of ensemble components, are selected on a validation set, which is standard model selection rather than circular reasoning. The self-citations to Peharz et al. (2019) and Stelzner et al. (2019) appear as background motivation for random SPNs and prior applications; they are not used as evidence for the paper's own improvement claims. The skeptical concern that Algorithm 4 may add residual links to product nodes or leaves and thereby violate decomposability is a correctness and model-validity issue, not a circularity issue; under the stated circularity criteria, no load-bearing step reduces by construction to its own inputs.

Assumptions & free parameters 5 free parameters · 3 assumptions · 1 invented entities

The central claim depends on hyperparameters (beta, gamma, k, n, EM iterations) that are set by hand or tuned on validation sets. The validity of the method rests on standard SPN structural constraints, on the unproven assumption that scope-matching additions preserve useful structure under joint EM, and on the comparability of external baselines. The only new entity is the residual link, which lacks independent falsifiable evidence beyond the paper's own likelihood comparisons.

free parameters (5)

beta = 0.6
Probability that the random feature split or instance clustering fails in ExtraSPN (Algorithm 3); set by hand.
mu range factor gamma = 5
mu is sampled from [1, |D|/gamma] for each ExtraSPN (Section 3); chosen by hand to control depth and runtime.
residual link ratio k = 0.1 or 0.2 (searched)
Maximum fraction of nodes in the target SPN to which residual links are added (Algorithm 4); selected on a validation set (Section 5.3).
ensemble size n = 10 (main results); 3, 5, 10 in Table 4
Number of ExtraSPNs combined; fixed by experiment design.
EM iterations = up to 1000
Optimization limit with early stopping based on likelihood variance (Section 5.1).

assumptions (3)

standard math SPNs with decomposable product nodes and complete sum nodes encode valid probability distributions.
Taken from Poon and Domingos 2011; used throughout to guarantee valid models (Section 2.1).
domain assumption Adding a marginalized, scope-matching node from another SPN as a new child of a sum node preserves the validity of the SPN.
Stated in Section 4: the scope of the two nodes is the same, so the networks remain complete; assumed without proof that joint EM optimization then produces a useful ensemble.
domain assumption The benchmark splits and preprocessing used for LearnSPN and ID-SPN baselines in [Rooshenas and Lowd, 2014] are identical to those in this paper's Table 1.
Baselines are not re-run; the paper compares against numbers reported in that earlier paper (Section 5.1, Table 2).

invented entities (1)

residual link
purpose: Connects a node in one SPN to a scope-matching (or marginalized) node in another SPN, adding the referenced node as an extra child of a sum node to refine density estimates.
The link is a new graph edge type introduced in Section 4. It has no predictive handle outside the paper's own experiments, and the empirical evidence consists of point estimates without error bars.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Random Sum-Product Forests with Residual Links." pith.science (2026). https://pith.science/paper/OR4Y23OR

@misc{pith2026190803250,
  author       = {Pith},
  title        = {Pith review of: Random Sum-Product Forests with Residual Links},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/OR4Y23OR}},
  note         = {Machine review of arXiv:1908.03250}
}

read the original abstract

Tractable yet expressive density estimators are a key building block of probabilistic machine learning. While sum-product networks (SPNs) offer attractive inference capabilities, obtaining structures large enough to fit complex, high-dimensional data has proven challenging. In this paper, we present random sum-product forests (RSPFs), an ensemble approach for mixing multiple randomly generated SPNs. We also introduce residual links, which reference specialized substructures of other component SPNs in order to leverage the context-specific knowledge encoded within them. Our empirical evidence demonstrates that RSPFs provide better performance than their individual components. Adding residual links improves the models further, allowing the resulting ResSPNs to be competitive with commonly used structure learning methods.

Figures

Figures reproduced from arXiv: 1908.03250 by the authors.

**Figure 2.** Illustration of Residual Sum-Product Learning. The blue [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

42 extracted references · 39 canonical work pages

[1]

Learning the structure of sum-product networks via an svd-based algorithm

[Adel et al., 2015] Tameem Adel, David Balduzzi, and Ali Ghodsi. Learning the structure of sum-product networks via an svd-based algorithm. In Proceedings of UAI,

work page 2015
[3]

Random forests

[Breiman, 2001] Leo Breiman. Random forests. Machine Learning, 45(1):5–32,

2001
[12]

Markov network structure learning: A randomized feature generation approach

[Haaren and Davis, 2012] Jan Van Haaren and Jesse Davis. Markov network structure learning: A randomized feature generation approach. In Proceedings of AAAI,

work page 2012
[14]

[Hoffman, 2017] Matthew D. Hoffman. Learning deep la- tent Gaussian models with Markov chain Monte Carlo. In Proceedings of ICML, pages 1510–1519,

work page 2017
[15]

Learning deep ResNet blocks sequentially using boosting theory

[Huang et al., 2018] Furong Huang, Jordan Ash, John Lang- ford, and Robert Schapire. Learning deep ResNet blocks sequentially using boosting theory. In Proceedings of ICML, pages 2058–2067,

work page 2018
[16]

Online algorithms for sum-product networks with continuous variables

[Jaini et al., 2016] Priyank Jaini, Abdullah Rashwan, Han Zhao, Yue Liu, Ershad Banijamali, Zhitang Chen, and Pas- cal Poupart. Online algorithms for sum-product networks with continuous variables. In Proceedings of PGMs,

work page 2016
[17]

Jaini, A

[Jaini et al., 2018] P. Jaini, A. Ghose, and P. Poupart. Prometheus: Directly learning acyclic directed graph structures for sum-product networks. In Proceedings of PGM,

work page 2018
[18]

[Kingma and Dhariwal, 2018] D. P. Kingma and P. Dhari- wal. Glow: Generative ﬂow with invertible 1x1 convolu- tions. In Proceedings of NIPS, pages 10236–10245,

work page 2018

Show all 42 references

[19]

[Kingma and Welling, 2014] D. P. Kingma and M. Welling. Auto-encoding variational Bayes. InProceedings of ICLR,

2014
[20]

The Neural Autoregressive Distribution Estima- tor

[Larochelle and Murray, 2011] Hugo Larochelle and Iain Murray. The Neural Autoregressive Distribution Estima- tor. In Proceedings of AISTATS, pages 29–37,

2011
[22]

Learning Markov network structure with decision trees

[Lowd and Davis, 2010] Daniel Lowd and Jesse Davis. Learning Markov network structure with decision trees. In Proceedings of IEEE ICDM, pages 334–343,

2010
[24]

Poisson sum-product networks: A deep architecture for tractable multivariate poisson distri- butions

[Molina et al., 2017] Alejandro Molina, Sriraam Natarajan, and Kristian Kersting. Poisson sum-product networks: A deep architecture for tractable multivariate poisson distri- butions. In Proceedings of AAAI,

2017
[25]

Mixed sum-product networks: A deep architecture for hybrid domains

[Molina et al., 2018] Alejandro Molina, Antonio Vergari, Nicola Di Mauro, Sriraam Natarajan, Floriana Esposito, and Kristian Kersting. Mixed sum-product networks: A deep architecture for hybrid domains. In Proceedings of AAAI, pages 3828–3835,

2018
[26]

Spﬂow: An easy and extensible library for deep prob- abilistic learning using sum-product networks

[Molina et al., 2019] Alejandro Molina, Antonio Ver- gari, Karl Stelzner, Robert Peharz, Pranav Subramani, Nicola Di Mauro, Pascal Poupart, and Kristian Kersting. Spﬂow: An easy and extensible library for deep prob- abilistic learning using sum-product networks. CoRR, abs/1901.03704,

2019 arXiv
[29]

Modeling speech with sum-product networks: Application to bandwidth exten- sion

[Peharz et al., 2014] Robert Peharz, Georg Kapeller, Pejman Mowlaee, and Franz Pernkopf. Modeling speech with sum-product networks: Application to bandwidth exten- sion. In Proceedings of ICASSP,

2014
[30]

Random Sum-Product Net- works: A Simple and Effective Approach to Probabilistic Deep Learning

[Peharz et al., 2019] Robert Peharz, Antonio Vergari, Karl Stelzner, Alejandro Molina, Martin Trapp, Kristian Kerst- ing, and Zoubin Ghahramani. Random Sum-Product Net- works: A Simple and Effective Approach to Probabilistic Deep Learning. Proceedings of UAI,

2019
[31]

Foundations of Sum-Product Networks for Probabilistic Modeling

[Peharz, 2015] Robert Peharz. Foundations of Sum-Product Networks for Probabilistic Modeling . PhD thesis, Graz University of Technology, SPSC,

2015
[32]

Sum-Product Networks: a New Deep Architecture

[Poon and Domingos, 2011] Hoifung Poon and Pedro Domingos. Sum-Product Networks: a New Deep Architecture. Proceedings of UAI,

2011
[33]

Merging strategies for sum-product networks: From trees to graphs

[Rahman and Gogate, 2016] Tahrima Rahman and Vibhav Gogate. Merging strategies for sum-product networks: From trees to graphs. In Proceedings of UAI,

2016
[34]

Discriminative non-parametric learning of arithmetic circuits

[Ramanan et al., 2019] Nandini Ramanan, Mayukh Das, Kristian Kersting, and Sriraam Natarajan. Discriminative non-parametric learning of arithmetic circuits. In Working Notes of the ICML Workshop on Tractable Probabilistic Models (TPM)

2019
[35]

Black box variational inference

[Ranganath et al., 2014] Rajesh Ranganath, Sean Gerrish, and David Blei. Black box variational inference. In Pro- ceedings of AISTATS, pages 814–822,

2014
[36]

Sum-product networks for structured pre- diction: Context-speciﬁc deep conditional random ﬁelds

[Ratajczak et al., 2014] Martin Ratajczak, S Tschiatschek, and F Pernkopf. Sum-product networks for structured pre- diction: Context-speciﬁc deep conditional random ﬁelds. Proceedings of the ICML Workshop on Learning Tractable Probabilistic Models, pages 1–10,

2014
[37]

[Rezende et al., 2014] D. J. Rezende, S. Mohamed, and D. Wierstra. Stochastic backpropagation and approximate inference in deep generative models. In Proceedings of ICML, pages 1278–1286,

2014
[38]

Learning Sum-Product Networks with Direct and Indirect Variable Interactions

[Rooshenas and Lowd, 2014] Amirmohammad Rooshenas and Daniel Lowd. Learning Sum-Product Networks with Direct and Indirect Variable Interactions. In Proceedings of ICML,

2014
[39]

Faster attend-infer-repeat with tractable probabilistic models

[Stelzner et al., 2019] Karl Stelzner, Robert Peharz, and Kristian Kersting. Faster attend-infer-repeat with tractable probabilistic models. In Proceedings of ICML,

2019
[40]

Wilber, and Serge J

[Veit et al., 2016] Andreas Veit, Michael J. Wilber, and Serge J. Belongie. Residual networks behave like ensem- bles of relatively shallow networks. In Proceedings of NIPS, pages 550–558,

2016
[41]

Simplifying, Regularizing and Strengthening Sum-Product Network Structure Learning

[Vergariet al., 2015] Antonio Vergari, Nicola Di Mauro, and Floriana Esposito. Simplifying, Regularizing and Strengthening Sum-Product Network Structure Learning. In Proceedings of ECML-PKDD,

2015
[42]

Modeling spatial layout for scene image understanding via a novel multiscale sum-product network

[Yuan et al., 2016] Zehuan Yuan, Hao Wang, Limin Wang, Tong Lu, Shivakumara Palaiahnakote, and Chew Lim Tan. Modeling spatial layout for scene image understanding via a novel multiscale sum-product network. Proceedings of Expert Systems with Applications, 63:231 – 240, 2016

2016
[1998]

Greedy Part-Wise Learning of Sum- Product Networks

[Peharz et al., 2013] Robert Peharz, Bernhard Geiger, and Franz Pernkopf. Greedy Part-Wise Learning of Sum- Product Networks. InProceedings of ECML-PKDD,

2013
[2001]

Efﬁcient principled learning of thin junction trees

[Chechetka and Guestrin, 2008] Anton Chechetka and Car- los Guestrin. Efﬁcient principled learning of thin junction trees. In Proceedings of NIPS, pages 273–280

2008
[2006]

[Goodfellow et al., 2014] I. J. Goodfellow, J. Pouget- Abadie, M. Mirza, B. Xu, D. Warde-Farley, S. Ozair, A. Courville, and Y . Bengio. Generative adversarial nets. In Proceedings of NIPS, pages 2672–2680,

2014
[2008]

Language modeling with Sum-Product Networks

[Cheng et al., 2014] Wei-Chen Cheng, Stanley Kok, Hoai Vu Pham, Hai Leong Chieu, and Kian Ming Adam Chai. Language modeling with Sum-Product Networks. In Proceedings of INTERSPEECH , pages 2098–2102,

2014
[2010]

Learning Markov networks with arithmetic circuits

[Lowd and Rooshenas, 2013] Daniel Lowd and Amirmo- hammad Rooshenas. Learning Markov networks with arithmetic circuits. In Proceedings of AISTATS , pages 406–414,

2013
[2011]

Liang, J

[Liang et al., 2017] Y . Liang, J. Bekker, and G. Van den Broeck. Learning the structure of probabilistic sentential decision diagrams. In Proceedings of UAI,

2017
[2012]

Deep residual learning for image recog- nition

[He et al., 2016] Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recog- nition. In Proceedings of CVPR, pages 770–778,

2016
[2013]

Extremely randomized trees

[Geurts et al., 2006] Pierre Geurts, Damien Ernst, and Louis Wehenkel. Extremely randomized trees. Machine Learn- ing, 63(1):3–42,

2006
[2014]

Di Mauro, A

[Di Mauro et al., 2017] N. Di Mauro, A. Vergari, T. Basile, and F. Esposito. Fast and accurate density estimation with extremely randomized cutset networks. In Proceedings of ECML/PKDD, pages 203–219,

2017
[2015]

Sum product networks for activity recognition

[Amer and Todorovic, 2015] Mohamed Amer and Sinisa Todorovic. Sum product networks for activity recognition. IEEE Transactions on Pattern Analysis and Machine In- telligence,

2015
[2016]

Learning the Structure of Sum-Product Networks

[Gens and Domingos, 2013] Robert Gens and Pedro Domin- gos. Learning the Structure of Sum-Product Networks. In Proceedings of the ICML, pages 873–880,

2013
[2017]

Ventola, and Antonio Vergari

[Di Mauro et al., 2018] Nicola Di Mauro, Floriana Esposito, Fabrizio G. Ventola, and Antonio Vergari. Sum-product network structure learning by efﬁcient product nodes dis- covery. Intelligenza Artiﬁciale, 12(2):143–159,

2018
[2018]

[Dinh et al., 2016] L. Dinh, J. Sohl-Dickstein, and S. Ben- gio. Density estimation using real NVP. arXiv preprint arXiv:1605.08803,

2016 arXiv
[2019]

Newman, Blake D.J., and C.J

[Newman et al., 1998] C.L. Newman, Blake D.J., and C.J. Merz. UCI repository of machine learning databases,

1998

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [1]

Learning the structure of sum-product networks via an svd-based algorithm

[Adel et al., 2015] Tameem Adel, David Balduzzi, and Ali Ghodsi. Learning the structure of sum-product networks via an svd-based algorithm. In Proceedings of UAI,

work page 2015

[2] [3]

Random forests

[Breiman, 2001] Leo Breiman. Random forests. Machine Learning, 45(1):5–32,

2001

[3] [12]

Markov network structure learning: A randomized feature generation approach

[Haaren and Davis, 2012] Jan Van Haaren and Jesse Davis. Markov network structure learning: A randomized feature generation approach. In Proceedings of AAAI,

work page 2012

[4] [14]

[Hoffman, 2017] Matthew D. Hoffman. Learning deep la- tent Gaussian models with Markov chain Monte Carlo. In Proceedings of ICML, pages 1510–1519,

work page 2017

[5] [15]

Learning deep ResNet blocks sequentially using boosting theory

[Huang et al., 2018] Furong Huang, Jordan Ash, John Lang- ford, and Robert Schapire. Learning deep ResNet blocks sequentially using boosting theory. In Proceedings of ICML, pages 2058–2067,

work page 2018

[6] [16]

Online algorithms for sum-product networks with continuous variables

[Jaini et al., 2016] Priyank Jaini, Abdullah Rashwan, Han Zhao, Yue Liu, Ershad Banijamali, Zhitang Chen, and Pas- cal Poupart. Online algorithms for sum-product networks with continuous variables. In Proceedings of PGMs,

work page 2016

[7] [17]

Jaini, A

[Jaini et al., 2018] P. Jaini, A. Ghose, and P. Poupart. Prometheus: Directly learning acyclic directed graph structures for sum-product networks. In Proceedings of PGM,

work page 2018

[8] [18]

[Kingma and Dhariwal, 2018] D. P. Kingma and P. Dhari- wal. Glow: Generative ﬂow with invertible 1x1 convolu- tions. In Proceedings of NIPS, pages 10236–10245,

work page 2018