Pith. sign in

REVIEW 3 major objections 5 minor 57 references

Scalable Probabilistic Matrix Factorization with Graph-Based Priors

T0 review · 3 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

Pith's one-line read Prune bad graph edges to make matrix factorization better and faster

desk verdict A practical graph-pruning step for matrix factorization that works in experiments, but the contested-edge story is better supported empirically than theoretically. read the letter →

arxiv 1908.09393 v2 pith:D4RCRGUS submitted 2019-08-25 cs.LG stat.ML

classification cs.LGstat.ML
keywords matrixfactorizationgraphside-informationcontestededgesgraphicallassoexpectationmaximizationcollaborativefilteringprobabilisticlarge-scalegraphs
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 noisy, irrelevant graph side-information hurts matrix-factorization prediction, and that the fix is to delete graph edges whose latent-feature covariance is negative, the contested edges. It introduces GPMF, a probabilistic model that places a graph-shaped Gaussian prior on the columns of the latent feature matrices, and GRAEM, an expectation-maximization algorithm that alternates a graph-regularized least-squares E-step with a cheap M-step that prunes contested edges by thresholding an estimated sample covariance matrix at zero. The paper argues this pruning is a graphical-lasso approximation and adds no computational complexity: the method stays linear in the number of non-zeros, and runtime shrinks as edges are removed. A reader should care because, if correct, the result means graph side-information can be made both more accurate and cheaper to use through one simple step that any graph-regularized matrix-completion method could adopt. Experiments report lower RMSE than the scalable baseline on Flixster, MovieLens 100k/20M, and Epinions, and full processing of a 300k-node, three-million-edge Yahoo Music graph in under ten minutes on a laptop.

What carries the argument

The load-bearing object is the adjacency update rule: an edge survives only if the expected sample covariance $E[S_D]$ of the latent features is at least the threshold $\tau$ (set to zero), otherwise it is removed as contested. The paper computes this expectation by decomposing each outer product into posterior covariance plus outer product of posterior means, then approximates the posterior covariance with an unbiased Monte Carlo estimate from K samples drawn via sparse Cholesky factorization of a block-diagonal approximation to the posterior precision; this column-wise independence assumption makes the Cholesky step linear in the graph size. The step is justified as a constrained graphical-lasso approximation by sign-consistency results saying that, for sparse and large problems, thresholding the sample covariance marks the same non-zeros as the graphical lasso would.

What would settle it

Take a synthetic 400 by 400 setup with a known true graph and a known fraction of planted contested edges, run GRAEM's zero-threshold M-step, and also run an explicit constrained graphical-lasso solve on the same expected sample covariance, then compare the two edge sets and the RMSE of each; if the sets diverge substantially under the paper's reported settings (7% observed entries, $\sigma^2 = 0.01$), the claimed equivalence is not what drives the accuracy gain.

Watch

Extended reading notes

Core claim

The central discovery is that a graph edge is worth keeping only if the latent features of its endpoints are positively correlated given the observed ratings; edges with negative estimated covariance actively mislead the factorization, and removing them helps. GPMF formalizes this by replacing the spherical priors of probabilistic matrix factorization with full-covariance Gaussian priors whose precision matrices are regularized Laplacians built from the graph, so the prior becomes a Gaussian Markov random field. Using EM, the M-step estimates the expected sample covariance of the latent feature columns and removes entries whose covariance is below $\tau=0$, which the authors identify as contested edges via sign-consistency results for the graphical lasso. The paper shows the E-step MAP objective equals the GRALS objective, giving convergence guarantees, and that the extra M-step is linear-cost; on real benchmarks the pruned graphs improve RMSE while using fewer edges, and on a densified 40-nearest-neighbour MovieLens20M graph GPMF improves while the fixed-graph baseline worsens.

Load-bearing premise

The rule that decides which edges are contested relies on the theorem that thresholding the sample covariance at zero marks exactly the edges the graphical lasso would set to zero; the paper sets the threshold to zero on real datasets without verifying the theorem's conditions, so if those conditions fail the deleted edges need not be the contested ones.

Editorial extensions

If this is right

  • On the reported benchmarks GPMF/GRAEM achieves RMSE 0.8857 on Flixster versus 0.9152 for GRALS, and 0.7887 on MovieLens20M (40NN) versus 0.7922, so pruning edges improves accuracy over keeping the full graph.
  • The method's complexity remains linear in the number of non-zeros; each removed edge shrinks the adjacency matrix, making the regularized least-squares iterations cheaper, and the paper demonstrates a 300k by 300k graph with 3 million edges updating in under ten minutes on a laptop.
  • The generated graph is reusable: feeding the pruned graph to KPMF preserves or improves accuracy (77% of Douban user edges, 65% of MovieLens 100k edges), suggesting contested-edge removal is a general preprocessing step for graph-regularized matrix completion.
  • Increasing graph density from 10 to 40 nearest neighbours degrades GRALS but lets GPMF keep improving, so noisy graphs can be rendered useful instead of discarded.
  • On synthetic data with 10% observed entries, GPMF nearly matches the accuracy of the true graph, while GRALS needs about 30% observations for comparable accuracy; when observation noise grows, GPMF degrades to at worst the corrupted-graph baseline.

Reading between the lines

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

  • Not stated in the paper: the zero threshold is a heuristic rather than a verified regime of the underlying equivalence theorem. One could stress-test it by replacing Equation (8) with an explicit constrained graphical-lasso solve on the same covariance and comparing both the edge sets and final RMSE; the paper's own synthetic numbers (only 31.7% to 44.3% of planted contested edges recovered) sugge
  • Not stated in the paper: the column-wise independence assumption that makes the posterior covariance tractable is also a modelling restriction, because a full inverse-Kronecker-sum covariance would couple the latent-feature columns. Extending the M-step to a low-rank plus diagonal correction could capture more signal without breaking linear complexity.
  • Not stated in the paper: because the M-step is written as an adjacency update, it is a model-selection rule for the Gaussian Markov random field graph itself, which connects the method to graph-learning problems and suggests an iterative or annealed version could trade graph sparsity against accuracy in a principled way.
  • A testable extension, absent from the paper, is to apply the pruned graph to graph-convolutional matrix-completion models, which the paper benchmarks but does not train on the updated graph; if the gain transfers, contested-edge removal is a model-agnostic graph-denoising step.
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 / 5 minor

Summary. The manuscript introduces GPMF, a probabilistic formulation of graph-regularised matrix factorization in which the precision matrices of the latent-feature priors are graph Laplacians, and GRAEM, an alternating algorithm that—after a PMF initialization—uses a thresholded posterior sample-covariance matrix to delete graph edges whose latent-feature correlation is negative (the 'contested edges'). The authors show in Lemma 1 that, conditional on a fixed graph, the MAP objective coincides with the GRALS objective. The central empirical claims are that pruning contested edges improves RMSE relative to GRALS on Flixster, MovieLens100k, Epinions, and MovieLens20M, and that the pruning step has linear cost, allowing a 300k-node/3M-edge graph to be processed in under ten minutes on a laptop.

Significance. If the identification step works as claimed, the contribution is practically valuable: a linear-time graph-pruning heuristic that improves accuracy and can be appended to GRALS, with a principled probabilistic derivation and experiments on standard benchmarks. The paper is also honest in reporting mixed results (Yahoo Music, KPMF with pruned graphs) and in acknowledging that the GLASSO approximation is only moderately successful in simulation. The main significance is limited, however, by the fact that the contested-edge identification—the mechanism behind the claimed accuracy gains—is not convincingly validated, and the claimed convergence guarantee does not follow from the cited tools.

major comments (3)
  1. [§3.3.3, Eq. (8)] The pruning rule in Eq. (8) thresholds E[S_D] at τ=0 and is justified by the sign-consistency and sparsity-equivalence results of Fattahi and Sojoudi. Those results are stated for a sufficiently large penalty τ with a highly sparse GLASSO solution (roughly 10N non-zeros). At τ=0 the GLASSO penalty vanishes, the solution is the dense inverse sample covariance, and the support-equivalence argument does not apply. The paper's own synthetic evaluation (Section 4.1) reports that only 31.7% of true contested edges are removed and 19% of true edges are wrongly removed at 7% observations, and the conclusion states that the GLASSO approximation is 'only moderately successful.' Because Eq. (8) is the entire M-step, the central claim that accuracy gains come from removing contested edges rests on an unvalidated approximation. I ask the authors to verify sign-consistency at the operating point, to compare against thresholding at positive τ where the cited equivalence has support, and to include a control such as random edge removal at the same rate to show the gains are not generic sparsification.
  2. [§3.3.1–3.3.4] The quantity thresholded in Eq. (8) is E[S_D] = (1/D) Σ_d (Σ_post_{U:d} + μ_post_{U:d} μ_post_{U:d}^T), a posterior second moment computed under a column-wise-independence approximation and MAP substitution. This is not an empirical sample covariance of i.i.d. draws from the model, so the cited GLASSO/thresholding equivalence—which is a statement about sample covariance matrices—does not directly apply to the object being thresholded. The authors should either derive the thresholding rule for posterior second moments under their approximate posterior, or explicitly treat Eq. (8) as a heuristic and validate it empirically; the current text moves between 'MLE' and 'highly-efficient approximation' without establishing which claim is being made.
  3. [Section 3, Algorithm 1] The abstract and Section 3 state that the EM formulation 'guarantees convergence' and that the method inherits the global convergence guarantees of GRALS via Xu and Yin. Those guarantees apply to block-coordinate minimization of a fixed biconvex objective. Here the M-step does not maximize Q (it heuristically changes the graph), and the graph change alters the regularizer in the subsequent E-step, so the objective is not fixed across iterations. The convergence claim is therefore unsupported as stated. Please state the precise convergence property that is guaranteed, or weaken the claim to empirical convergence.
minor comments (5)
  1. [Section 4.2, Table 1] The sentence 'we get the same accuracy for Flixster with almost halved edges' is accurate from the table, but the accompanying claim that the pruned graph improves 'arbitrary algorithms' is contradicted by the MovieLens100k KPMF result (0.9336 vs 0.9374).
  2. [Section 2, Eq. (4)] The notation switches between Λ_U and L_U^+ without consistently naming the regularization strength; in Eq. (4) the factor σ^2/2 appears without explaining that the posterior has been multiplied by σ^2.
  3. [Section 4.1, Figure 2] The text says 'at worst GPMF is only as bad as using the original corrupted graph,' but at high noise PMF appears competitive, and the comparison point is not clearly defined; please clarify.
  4. [Section 4.2, Figure 3] The proportion of remaining edges is reported in figure titles, but for MovieLens20M the caption says '40NN graph' while the text reports 10/20/40-NN results; it would help to state all edge-retention rates in one table.
  5. [Throughout] Several typos should be corrected: 'effect' for 'affect' (Section 1), 'helf' for 'half' (Section 4.2), 'samge' for 'same' (Appendix B.2), and inconsistent verb tense in Section 4.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: GPMF's graph update is an empirical mechanism evaluated on held-out entries, not a renamed input or self-citation-driven prediction.

full rationale

The derivation chain is not circular. GPMF defines a generative model whose MAP objective, under a regularized-Laplacian precision matrix, is shown in Lemma 1 to coincide with GRALS; this is an independent algebraic reduction. The M-step derives the precision MLE as the inverse of an expected sample covariance matrix (Eq. 6), and the contested-edge removal rule (Eq. 8) thresholds that expected covariance at tau=0. The GLASSO equivalence used to justify the thresholding is imported from external results (Fattahi and Sojoudi, Zhang et al., Mazumder and Hastie), not from the authors' own prior work, and it is not used as an unexamined premise of the paper's own method. The claimed accuracy gains are evaluated on held-out entries against external baselines on Flixster, Douban, MovieLens, Epinions, Yahoo Music, and MovieLens20M, so the result is falsifiable rather than forced by construction. The self-citations present (KBMF, Nguyen and Mamitsuka, Zheng et al.) are literature or baseline references and do not carry the contested-edge argument. The paper itself concedes that the GLASSO approximation is only moderately successful in simulated data, which is a robustness limitation rather than circularity. No fitted parameter is relabeled as a prediction, and no equation reduces to its own input; hence the appropriate circularity score is 0.

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

The main free parameters are the threshold tau and the tuned hyperparameters (D, sigma^2, lambdas, K). The load-bearing axioms are the GMRF/Laplacian-to-precision mapping, the sign-consistency assumption that connects negative sample covariances to removable edges, and the ad hoc column-wise independence approximation for the posterior covariance. The paper introduces no new physical or mathematical entities beyond the concept of a contested edge, which is a modeling construct rather than an invented entity.

free parameters (5)
  • threshold tau = 0
    Edges with expected covariance less than tau are removed; set to zero in all experiments, but can be increased for sparser graphs.
  • latent dimension D = 10 (20 for Yahoo Music)
    Tuned per dataset in Table 2.
  • observation noise sigma^2 = 0.05 to 10 depending on dataset
    Tuned per dataset in Table 2.
  • regularization strengths lambda_L, lambda_U, lambda_V = Various values in Table 2
    Tuned per dataset; control the Laplacian and feature priors.
  • posterior sample count K = not reported
    Number of samples used to estimate the posterior covariance in Section 3.3.4 is not specified, an implementation detail affecting the M-step.
assumptions (5)
  • standard math Precision matrix zero pattern equals graph conditional independence (GMRF property).
    Section 2.1; standard result in GMRF theory.
  • standard math Regularized Laplacian L+ = L + gamma I is a valid precision matrix with the same zero pattern as the graph.
    Section 2.2; cited to Dong et al. 2016, Egilmez et al. 2016 and others.
  • domain assumption Sign-consistency: non-zero entries of the GLASSO solution have signs opposite to the corresponding sample covariance entries, so negative sample covariances identify removable edges.
    Section 3.3.3 states 'Assuming sign-consistency'; the cited equivalence results (Fattahi and Sojoudi 2019) hold for sufficiently large tau and roughly 10N non-zeros, but the paper sets tau=0.
  • ad hoc to paper Column-wise independence of the posterior covariance: off-diagonal blocks of C in the Kronecker sum are set to zero, making the posterior precision block-diagonal.
    Section 3.3.4, Equation (9); an approximation introduced solely for scalability, not independently validated.
  • ad hoc to paper Sparsity-structure equivalence between the thresholded sample covariance matrix and the GLASSO solution holds at the operating point tau=0.
    Section 3.3.3 relies on this to justify Equation (8); the cited theory requires substantial sparsity and a positive penalty.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Scalable Probabilistic Matrix Factorization with Graph-Based Priors." pith.science (2026). https://pith.science/paper/D4RCRGUS

@misc{pith2026190809393,
  author       = {Pith},
  title        = {Pith review of: Scalable Probabilistic Matrix Factorization with Graph-Based Priors},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/D4RCRGUS}},
  note         = {Machine review of arXiv:1908.09393}
}
abstract

In matrix factorization, available graph side-information may not be well suited for the matrix completion problem, having edges that disagree with the latent-feature relations learnt from the incomplete data matrix. We show that removing these $\textit{contested}$ edges improves prediction accuracy and scalability. We identify the contested edges through a highly-efficient graphical lasso approximation. The identification and removal of contested edges adds no computational complexity to state-of-the-art graph-regularized matrix factorization, remaining linear with respect to the number of non-zeros. Computational load even decreases proportional to the number of edges removed. Formulating a probabilistic generative model and using expectation maximization to extend graph-regularised alternating least squares (GRALS) guarantees convergence. Rich simulated experiments illustrate the desired properties of the resulting algorithm. On real data experiments we demonstrate improved prediction accuracy with fewer graph edges (empirical evidence that graph side-information is often inaccurate). A 300 thousand dimensional graph with three million edges (Yahoo music side-information) can be analyzed in under ten minutes on a standard laptop computer demonstrating the efficiency of our graph update.

Figures

Figures reproduced from arXiv: 1908.09393 by the authors.

Figure 1
Figure 1. An illustrative movie recommendation problem. [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Synthetic data experiments with respect to the dimension size O(N + M), constrained SCM computation and thresholding, O(nz(AU ) + nz(AV )) both converge in one time step. Line 4 uses GRALS with the sparsified graphs: O(nz(Ω) + nz(A + U ) + nz(A + V )) per CG iteration. Line 4 is initialised with U,V values from the PMF run, largely reducing the number of iterations required. Our algorithm remains linear with respect… view at source ↗
Figure 3
Figure 3. Convergence time; vertical lines show start and end of M-step. c) [PITH_FULL_IMAGE:figures/full_fig_p012_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Convergence time on MovieLens 100k. We provide an updated graph [PITH_FULL_IMAGE:figures/full_fig_p025_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

57 extracted references · 52 canonical work pages

  1. [1]

    Managing uncertainty in social networks

    Eytan Adar and Christopher Re. Managing uncertainty in social networks. IEEE Data Eng. Bull., 30 0 (2): 0 15--22, 2007

  2. [2]

    Binary rating estimation with graph side information

    Kwangjun Ahn, Kangwook Lee, Hyunseung Cha, and Changho Suh. Binary rating estimation with graph side information. In Advances in Neural Information Processing Systems, pages 4272--4283, 2018

  3. [3]

    Predicting protein complex membership using probabilistic network reliability

    Saurabh Asthana, Oliver D King, Francis D Gibbons, and Frederick P Roth. Predicting protein complex membership using probabilistic network reliability. Genome research, 14 0 (6): 0 1170--1175, 2004

  4. [4]

    Graph convolutional matrix completion

    Rianne van den Berg, Thomas N Kipf, and Max Welling. Graph convolutional matrix completion. arXiv preprint arXiv:1706.02263, 2017

  5. [5]

    Christopher M. Bishop. Pattern Recognition and Machine Learning. Springer, 2006

  6. [6]

    Metagenes and molecular pattern discovery using matrix factorization

    Jean-Philippe Brunet, Pablo Tamayo, Todd R Golub, and Jill P Mesirov. Metagenes and molecular pattern discovery using matrix factorization. Proceedings of the national academy of sciences, 101 0 (12): 0 4164--4169, 2004

  7. [7]

    Graph regularized nonnegative matrix factorization for data representation

    Deng Cai, Xiaofei He, Jiawei Han, and Thomas S Huang. Graph regularized nonnegative matrix factorization for data representation. IEEE Transactions on Pattern Analysis and Machine Intelligence, 33 0 (8): 0 1548--1560, 2011

  8. [8]

    Matrix completion with noise

    Emmanuel J Candes and Yaniv Plan. Matrix completion with noise. Proceedings of the IEEE, 98 0 (6): 0 925--936, 2010

Show all 57 references
  1. [9]

    Exact matrix completion via convex optimization

    Emmanuel J Cand \`e s and Benjamin Recht. Exact matrix completion via convex optimization. Foundations of Computational mathematics, 9 0 (6): 0 717, 2009

  2. [10]

    Matrix completion with noisy side information

    Kai-Yang Chiang, Cho-Jui Hsieh, and Inderjit S Dhillon. Matrix completion with noisy side information. In C. Cortes, N. D. Lawrence, D. D. Lee, M. Sugiyama, and R. Garnett, editors, Advances in Neural Information Processing Systems 28, pages 3447--3455. Curran Associates, Inc....

  3. [11]

    Dhillon, and Cho-Jui Hsieh

    Kai-Yang Chiang, Inderjit S. Dhillon, and Cho-Jui Hsieh. Using side information to reliably learn low-rank matrices from missing and corrupted observations. Journal of Machine Learning Research, 19 0 (76): 0 1--35, 2018. URL http://jmlr.org/papers/v19/17-112.html

  4. [12]

    An overview of low-rank matrix recovery from incomplete observations

    Mark A Davenport and Justin Romberg. An overview of low-rank matrix recovery from incomplete observations. IEEE Journal of Selected Topics in Signal Processing, 10 0 (4): 0 608--622, 2016

  5. [13]

    A column approximate minimum degree ordering algorithm

    Timothy A Davis, John R Gilbert, Stefan I Larimore, and Esmond G Ng. A column approximate minimum degree ordering algorithm. ACM Transactions on Mathematical Software (TOMS), 30 0 (3): 0 353--376, 2004

  6. [14]

    Learning laplacian matrix in smooth graph signal representations

    Xiaowen Dong, Dorina Thanou, Pascal Frossard, and Pierre Vandergheynst. Learning laplacian matrix in smooth graph signal representations. IEEE Transactions on Signal Processing, 64 0 (23): 0 6160--6173, 2016

  7. [15]

    The yahoo! music dataset and kdd-cup'11

    Gideon Dror, Noam Koenigstein, Yehuda Koren, and Markus Weimer. The yahoo! music dataset and kdd-cup'11. In Proceedings of the 2011 International Conference on KDD Cup 2011-Volume 18, pages 3--18. JMLR. org, 2011

  8. [16]

    Graph learning with laplacian constraints: Modeling attractive gaussian markov random fields

    Hilmi E Egilmez, Eduardo Pavez, and Antonio Ortega. Graph learning with laplacian constraints: Modeling attractive gaussian markov random fields. In Signals, Systems and Computers, 2016 50th Asilomar Conference on, pages 1470--1474. IEEE, 2016

  9. [17]

    Graph learning from data under laplacian and structural constraints

    Hilmi E Egilmez, Eduardo Pavez, and Antonio Ortega. Graph learning from data under laplacian and structural constraints. IEEE Journal of Selected Topics in Signal Processing, 11 0 (6): 0 825--841, 2017

  10. [18]

    Graphical lasso and thresholding: Equivalence and closed-form solutions

    Salar Fattahi and Somayeh Sojoudi. Graphical lasso and thresholding: Equivalence and closed-form solutions. arXiv preprint arXiv:1708.09479, 2017

  11. [19]

    Graphical lasso and thresholding: equivalence and closed-form solutions

    Salar Fattahi and Somayeh Sojoudi. Graphical lasso and thresholding: equivalence and closed-form solutions. The Journal of Machine Learning Research, 20 0 (1): 0 364--407, 2019

  12. [20]

    Kernelized bayesian matrix factorization

    Mehmet G \"o nen, Suleiman Khan, and Samuel Kaski. Kernelized bayesian matrix factorization. In International Conference on Machine Learning, pages 864--872, 2013

  13. [21]

    Pathway graphical lasso

    Maxim Grechkin, Maryam Fazel, Daniela Witten, and Su-In Lee. Pathway graphical lasso. In Twenty-Ninth AAAI Conference on Artificial Intelligence, 2015

  14. [22]

    A systemic analysis of link prediction in social network

    Sogol Haghani and Mohammad Reza Keyvanpour. A systemic analysis of link prediction in social network. Artificial Intelligence Review, pages 1--35, 2017

  15. [23]

    The movielens datasets: History and context

    F Maxwell Harper and Joseph A Konstan. The movielens datasets: History and context. Transactions on Interactive Intelligent Systems, 5 0 (4), 2015

  16. [24]

    Deep models of interactions across sets

    Jason Hartford, Devon R Graham, Kevin Leyton-Brown, and Siamak Ravanbakhsh. Deep models of interactions across sets. arXiv preprint arXiv:1803.02879, 2018

  17. [25]

    Hastie, R

    T. Hastie, R. Tibshirani, and J. Friedman. The Elements of Statistical Learning: Data Mining, Inference, and Prediction, Second Edition. Springer Series in Statistics. Springer New York, 2009

  18. [26]

    The future of computational chemogenomics

    Edgar Jacoby and JB Brown. The future of computational chemogenomics. In Computational Chemogenomics, pages 425--450. Springer, 2018

  19. [27]

    The bigraphical lasso

    Alfredo Kalaitzis, John Lafferty, Neil Lawrence, and Shuheng Zhou. The bigraphical lasso. In International Conference on Machine Learning, pages 1229--1237, 2013

  20. [28]

    Matrix factorization techniques for recommender systems

    Yehuda Koren, Robert Bell, and Chris Volinsky. Matrix factorization techniques for recommender systems. Computer, 2009

  21. [29]

    Lauritzen

    S.L. Lauritzen. Graphical Models. Oxford science publications. Clarendon Press, 1996

  22. [30]

    Lrbm: A restricted boltzmann machine based approach for representation learning on linked data

    Kang Li, Jing Gao, Suxin Guo, Nan Du, Xiaoyi Li, and Aidong Zhang. Lrbm: A restricted boltzmann machine based approach for representation learning on linked data. In 2014 IEEE International Conference on Data Mining, pages 300--309. IEEE, 2014

  23. [31]

    Bayesian regularization via graph laplacian

    Fei Liu, Sounak Chakraborty, Fan Li, Yan Liu, Aurelie C Lozano, et al. Bayesian regularization via graph laplacian. Bayesian Analysis, 9 0 (2): 0 449--474, 2014

  24. [32]

    Graph summarization methods and applications: A survey

    Yike Liu, Tara Safavi, Abhilash Dighe, and Danai Koutra. Graph summarization methods and applications: A survey. ACM Computing Surveys (CSUR), 51 0 (3): 0 62, 2018

  25. [33]

    Recommender systems with social regularization

    Hao Ma, Dengyong Zhou, Chao Liu, Michael R Lyu, and Irwin King. Recommender systems with social regularization. In Proceedings of the fourth ACM international conference on Web search and data mining, pages 287--296. ACM, 2011

  26. [34]

    The graphical lasso: New insights and alternatives

    Rahul Mazumder and Trevor Hastie. The graphical lasso: New insights and alternatives. Electronic journal of statistics, 6: 0 2125, 2012

  27. [35]

    Birds of a feather: Homophily in social networks

    Miller McPherson, Lynn Smith-Lovin, and James M Cook. Birds of a feather: Homophily in social networks. Annual review of sociology, 27 0 (1): 0 415--444, 2001

  28. [36]

    A review on matrix factorization techniques in recommender systems

    Rachana Mehta and Keyur Rana. A review on matrix factorization techniques in recommender systems. In 2017 2nd International Conference on Communication Systems, Computing and IT Applications (CSCITA), pages 269--274. IEEE, 2017

  29. [37]

    Probabilistic matrix factorization

    Andriy Mnih and Ruslan R Salakhutdinov. Probabilistic matrix factorization. In Advances in neural information processing systems, pages 1257--1264, 2008

  30. [38]

    Geometric matrix completion with recurrent multi-graph neural networks

    Federico Monti, Michael Bronstein, and Xavier Bresson. Geometric matrix completion with recurrent multi-graph neural networks. In Advances in Neural Information Processing Systems, pages 3697--3707, 2017

  31. [39]

    Latent feature kernels for link prediction on sparse graphs

    Canh Hao Nguyen and Hiroshi Mamitsuka. Latent feature kernels for link prediction on sparse graphs. IEEE transactions on neural networks and learning systems, 23 0 (11): 0 1793--1804, 2012

  32. [40]

    Collaborative filtering with graph information: Consistency and scalable methods

    Nikhil Rao, Hsiang-Fu Yu, Pradeep K Ravikumar, and Inderjit S Dhillon. Collaborative filtering with graph information: Consistency and scalable methods. In Advances in neural information processing systems, pages 2107--2115, 2015

  33. [41]

    Rue and L

    H. Rue and L. Held. Gaussian Markov Random Fields: Theory and Applications. CRC Press, 2005

  34. [42]

    Optimizing parallel collaborative filtering approaches for improving recommendation systems performance

    Christos Sardianos, Grigorios Ballas Papadatos, and Iraklis Varlamis. Optimizing parallel collaborative filtering approaches for improving recommendation systems performance. Information, 10 0 (5): 0 155, 2019

  35. [43]

    On the kronecker product

    Kathrin Schacke. On the kronecker product. Master's thesis, University of Waterloo, 2004

  36. [44]

    Yes, there is a correlation:-from social networks to personal behavior on the web

    Parag Singla and Matthew Richardson. Yes, there is a correlation:-from social networks to personal behavior on the web. In Proceedings of the 17th international conference on World Wide Web, pages 655--664. ACM, 2008

  37. [45]

    Equivalence of graphical lasso and thresholding for sparse graphs

    Somayeh Sojoudi. Equivalence of graphical lasso and thresholding for sparse graphs. The Journal of Machine Learning Research, 17 0 (1): 0 3943--3963, 2016 a

  38. [46]

    Graphical lasso and thresholding: Conditions for equivalence

    Somayeh Sojoudi. Graphical lasso and thresholding: Conditions for equivalence. In Decision and Control (CDC), pages 7042--7048. IEEE, 2016 b

  39. [47]

    Enter the matrix: factorization uncovers knowledge from omics

    Genevieve L Stein-O’Brien, Raman Arora, Aedin C Culhane, Alexander V Favorov, Lana X Garmire, Casey S Greene, Loyal A Goff, Yifeng Li, Aloune Ngom, Michael F Ochs, et al. Enter the matrix: factorization uncovers knowledge from omics. Trends in Genetics, 2018

  40. [48]

    mtrust: discerning multi-faceted trust in a connected world

    Jiliang Tang, Huiji Gao, and Huan Liu. mtrust: discerning multi-faceted trust in a connected world. In Proceedings of the fifth ACM international conference on Web search and data mining, pages 93--102. ACM, 2012

  41. [49]

    A block coordinate descent method for regularized multiconvex optimization with applications to nonnegative tensor factorization and completion

    Yangyang Xu and Wotao Yin. A block coordinate descent method for regularized multiconvex optimization with applications to nonnegative tensor factorization and completion. SIAM Journal on imaging sciences, 6 0 (3): 0 1758--1789, 2013

  42. [50]

    Depth image inpainting: Improving low rank matrix completion with low gradient regularization

    Hongyang Xue, Shengming Zhang, and Deng Cai. Depth image inpainting: Improving low rank matrix completion with low gradient regularization. IEEE Transactions on Image Processing, 26 0 (9): 0 4311--4320, 2017

  43. [51]

    Convolutional geometric matrix completion

    Kai-Lang Yao and Wu-Jun Li. Convolutional geometric matrix completion. arXiv preprint arXiv:1803.00754, 2018

  44. [52]

    Gene prioritization using bayesian matrix factorization with genomic and phenotypic side information

    Pooya Zakeri, Jaak Simm, Adam Arany, Sarah ElShal, and Yves Moreau. Gene prioritization using bayesian matrix factorization with genomic and phenotypic side information. Bioinformatics, 34 0 (13): 0 i447--i456, 2018

  45. [53]

    Large-scale sparse inverse covariance estimation via thresholding and max-det matrix completion

    Richard Zhang, Salar Fattahi, and Somayeh Sojoudi. Large-scale sparse inverse covariance estimation via thresholding and max-det matrix completion. In Proceedings of the 35th International Conference on Machine Learning, pages 5766--5775. PMLR, 2018

  46. [54]

    Leveraging node attributes for incomplete relational data

    He Zhao, Lan Du, and Wray Buntine. Leveraging node attributes for incomplete relational data. In Proceedings of the 34th International Conference on Machine Learning-Volume 70, pages 4072--4081. JMLR. org, 2017

  47. [55]

    Zhao , L

    Z. Zhao , L. Zhang , X. He , and W. Ng . Expert finding for question answering via graph regularized matrix completion. IEEE Transactions on Knowledge and Data Engineering, 27 0 (4): 0 993--1004, April 2015. ISSN 1041-4347. doi:10.1109/TKDE.2014.2356461

  48. [56]

    Collaborative matrix factorization with multiple similarities for predicting drug-target interactions

    Xiaodong Zheng, Hao Ding, Hiroshi Mamitsuka, and Shanfeng Zhu. Collaborative matrix factorization with multiple similarities for predicting drug-target interactions. In Proceedings of the 19th ACM SIGKDD international conference on Knowledge discovery and data mining, pages 10...

  49. [57]

    Kernelized probabilistic matrix factorization: Exploiting graphs and side information

    Tinghui Zhou, Hanhuai Shan, Arindam Banerjee, and Guillermo Sapiro. Kernelized probabilistic matrix factorization: Exploiting graphs and side information. In Proceedings of the 2012 SIAM International Conference on Data Mining, pages 403--414. SIAM, 2012

Pith tools

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