Pith. sign in

REVIEW 3 major objections 4 minor 2 cited by

Interpretable Bayesian Tensor Network Kernel Machines with Automatic Rank and Feature Selection

T0 review · 3 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read This paper claims that a mean-field variational approximation over tensor-network factors yields a Bayesian alternating least-squares algorithm whose computational complexity matches the deterministic version, so predictive uncertainty is…

desk verdict Solid Bayesian TN kernel machine with ARD, but the predictive variance is a first-order approximation and the zero-cost claim is overstated. read the letter →

arxiv 2507.11136 v1 pith:DWGCLGPH submitted 2025-07-15 stat.ML cs.LG

Afra Kilic , Kim Batselier This is my paper · ORCID
classification stat.MLcs.LG
keywords tensornetworkkernelmachinesvariationalinferenceautomaticrelevancedeterminationrankselectionfeatureuncertaintyquantificationBayesianalternatingleastsquaresKroneckerproductfeatures
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

Tensor-network kernel machines represent the weight vector of a kernel model as a low-rank tensor decomposition, which makes learning fast but leaves two problems: the weights are point estimates with no uncertainty, and the tensor rank and feature dimensions must be tuned by hand. This paper proposes a fully Bayesian version in which every factor matrix carries a Gaussian prior and the precisions of those priors are themselves random variables with Gamma hyperpriors. The paper's central claim is that a mean-field variational approximation on the tensor-network factors produces a Bayesian alternating least-squares algorithm with the same computational complexity as its deterministic counterpart, so uncertainty quantification comes at no additional asymptotic cost. Under this model, the hierarchical priors shrink entire rank components and irrelevant feature rows toward zero, so tensor rank and feature dimensions are inferred from data. Experiments on synthetic and real datasets are presented as evidence that the method recovers ground-truth complexity, gives calibrated predictive intervals, and scales to a dataset with 45,222 samples and 96 features.

What carries the argument

The machinery is a conjugate mean-field variational Bayes algorithm over a CP-decomposed weight tensor. Factor matrices $W^{(d)} \in \mathbb{R}^{M_d \times R}$ are assigned zero-mean Gaussian priors with precision $\Lambda_R \otimes \Lambda_{M_d}$; the shared precision vector $\lambda_R$ prunes whole rank columns, $\lambda_{M_d}$ prunes feature rows, and Gamma hyperpriors on both implement automatic relevance determination. The posterior covariance $\Sigma^{(d)}$ in Eq. (21) is the load-bearing object: its diagonal generates the variance terms feeding the hyperparameter and predictive updates, and the reshaping identities in Theorems 4 and 6 express second-moment terms as Khatri-Rao and Hadamard products of already-computed matrices. This preserves the exact ALS structure in the mean update while adding covariance terms, and integrating out the noise precision yields a Student's $t$ predictive distribution.

What would settle it

Take a small tensor regression problem with strongly correlated true factor columns, run the variational algorithm and an MCMC sampler on the same model, and compare predictive interval coverage and selected rank; if the variational intervals under-cover systematically or the selected rank diverges from the MCMC posterior, the cost-free uncertainty claim fails.

Watch

Extended reading notes

Core claim

The central discovery is a conjugation structure: when the weight vector of a kernel machine is CP-decomposed as $w = \sum_r w_r^{(1)} \otimes \cdots \otimes w_r^{(D)}$, placing zero-mean Gaussian priors on the factor matrices with Kronecker precisions $\Lambda_R \otimes \Lambda_{M_d}$, and Gamma hyperpriors on those precisions, makes the variational posterior updates close under the mean-field factorization. The update for each factor matrix takes the form of an alternating least-squares step, with the posterior mean given by the same Khatri-Rao and Hadamard design matrices used by deterministic ALS, and the posterior covariance $\Sigma^{(d)}$ carrying the additional uncertainty terms. Those covariance terms propagate into Gamma updates for the shared column-precision $\lambda_R$ and the per-mode row-precision $\lambda_{M_d}$, giving automatic rank and feature selection, and into a Student's $t$ predictive distribution for new targets. The paper's headline result is that all of this happens at the same $O\left(\sum_d N M_d^2 R^2 + M_d^3 R^3\right)$ per-iteration complexity as deterministic ALS, so the Bayesian machinery is not paid for in asymptotic runtime.

Load-bearing premise

The load-bearing premise is that the mean-field approximation, which treats all factor matrices as independent in the posterior, is accurate enough that the Student's $t$ predictive intervals and the automatically selected tensor ranks remain well calibrated.

Editorial extensions

If this is right

  • Uncertainty quantification becomes a free add-on to tensor-network kernel machines, with the same asymptotic per-iteration cost as deterministic ALS.
  • Tensor rank and feature dimensions no longer need manual cross-validation; the hierarchical priors shrink unnecessary components to zero during training.
  • The per-mode precision vectors act as feature-relevance scores, so the trained model doubles as an interpretability tool.
  • The method scales to large high-dimensional datasets: in the paper's experiments, it handles a 45,222-sample, 96-feature classification task and matches or improves on Gaussian-process RMSE and NLL on several benchmarks.
  • A tensor-train variant follows from the same construction by giving each TT rank its own precision vector, as Section 3.5 sketches.

Reading between the lines

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

  • Because the posterior covariance is already computed for every factor matrix, an active-learning or Bayesian-optimization extension could query points using the predictive variance at essentially no additional per-iteration cost; the paper does not pursue this use.
  • The negative correlation between the $\lambda_R$ and $\lambda_{M_d}$ updates means a heavily penalized rank component can be offset by relaxing row precisions, a compensation mechanism that could be isolated by ablating either hyperprior.
  • The same ARD-style hierarchical prior should transfer to tensor-train cores, and a full TT implementation would be a direct test of whether automatic multi-rank inference behaves as well as the CPD case; the paper only sketches this extension.
  • If the no-extra-cost claim survives at scale, deterministic tensor-kernel pipelines in production could be upgraded to output calibrated intervals and feature relevance without changing their training-time complexity; this is an inference from the paper's result, not a claim it makes.
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 BTN-Kernel Machines, a variational Bayesian extension of tensor-network kernel machines. A CPD (or TT) weight tensor is assigned a hierarchical sparsity-inducing prior: Gamma-distributed precision parameters act on the rank columns and on the feature rows of each factor matrix, with the aim of automatically inferring tensor rank and feature dimensions. Variational mean-field inference yields closed-form Gaussian posteriors for the factor matrices and Gamma posteriors for the precision hyperparameters, leading to alternating updates that the authors claim have the same asymptotic cost as deterministic ALS. Predictive uncertainty is summarized by a Student-t distribution. Experiments on synthetic data and several UCI data sets compare the method with GP, T-KRR, and SP-BTN in terms of RMSE, negative log-likelihood, effective rank, and effective feature dimension.

Significance. The paper addresses a relevant gap: deterministic tensor-network kernel machines do not provide posterior uncertainty, and existing probabilistic tensor methods often rely on restrictive diagonal covariances or manual complexity tuning. The core variational update equations, Theorems 3–6, and the moment formulas in the appendices are derived in detail and are largely internally consistent; the authors also release code for reproducing the experiments. If the predictive uncertainty were shown to be calibrated, the method would be a useful scalable Bayesian kernel machine with automatic complexity control. However, the two headline claims—uncertainty quantification at no extra cost and automatic feature-dimension inference—are currently supported only under additional approximations and threshold-dependent heuristics that are not fully acknowledged in the main text.

major comments (3)
  1. [Section 3.3 / Appendix 11] The predictive distribution in Eq. (43) is not an exact marginalization under the variational posterior. The appendix integrates over vec(W^(1)) while treating g^(d)(x_n) as fixed, but g^(d)(x_n) contains the remaining random factor matrices W^(k), k≠1; sequential integration of this kind is invalid. The resulting expression is a first-order delta-method approximation in which the predicted variance is built from terms g^(d)(x_n)^T Σ^(d) g^(d)(x_n) evaluated at posterior means. The exact second moment under the mean-field posterior is E_q[f^2] = Σ_{r,s} ∏_d (μ_{d,r} μ_{d,s} + C_{d,rs}), with μ_{d,r} = φ^(d)^T \tilde w_r^(d) and C_{d,rs} = φ^(d)^T Cov(w_r^(d), w_s^(d)) φ^(d); Eq. (43) omits products of two or more posterior covariances across different factors. The sign of the omitted remainder is not controlled, so the predictive variance can be either too large or too small, and the discussion in the paper does not establish calibration. The Student-t form is also itself an approximation after integrating out τ. This issue is load-bearing because the paper's central claim is uncertainty quantification at no extra cost; please restate the derivation as a first-order approximation and provide a quantitative assessment of its bias, e.g., by comparing against Monte Carlo moments from q on small-N examples.
  2. [Section 4.1 / Algorithm 1] The claimed automatic inference of feature dimensions M_d is not literally achieved. Algorithm 1 only prunes zero columns (rank components); rows of the factor matrices are retained throughout, and the paper itself states in Section 4.1 that strongly penalized rows become very small but not exactly zero. The experiments therefore measure M_eff using an ad hoc 0.25% variance threshold (Section 4.3), and the reported effective feature dimension depends on that threshold. This is not the same as automatically inferring a discrete feature dimension. Please either implement row pruning or explicitly redefine the claim as soft feature selection with a threshold-dependent effective dimension, and state the threshold's influence on the results.
  3. [Section 3.4] The computational-complexity claim 'same computational complexity as its deterministic counterpart' is a flop-count statement only. Maintaining the full covariance matrices Σ^(d) requires O(M_d^2 R^2) memory per mode, whereas deterministic ALS stores only O(M_d R) parameters per mode; the lower-bound evaluation in Eq. (42) and the truncation checks also add overhead. For large M_d and R the memory and constant-factor differences are substantial, so the abstract's phrase 'uncertainty quantification at no extra computational cost' overstates the situation. Please qualify the claim as asymptotic flop complexity and report memory complexity separately.
minor comments (4)
  1. [Eq. (43)] The predictive mean is written as \tilde y_i = ⊛_{d=1}^D \tilde W^(d) φ_i^(d), which is an R-vector under the paper's Hadamard-product notation, not a scalar. The scalar predictive mean should be 1_R^T (⊛_{d=1}^D \tilde W^(d) φ_i^(d)), or an explicit summation over rank components should be introduced.
  2. [Section 3.2.5] The text calls c0=d0=g0=h0=10^{-6} 'uninformative', but the experiments use c0=10^{-5} and d0=10^{-6} for the rank hyperparameters. Please align the description and the experimental settings, or explain why this difference is immaterial.
  3. [Table 4] SP-BTN results are taken from the original paper rather than measured under the same train/test splits as the other methods; this should be stated more prominently as a limitation of the comparison, not only in the experimental setup paragraph.
  4. [Appendix 8, Eq. (55)] The final expression in the proof of Theorem 6 uses ⊛_{k≠d} where the context requires ⊛_{d=1}^D; the index is inconsistent with the previous line and with Eq. (40).

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the Bayesian ALS updates and predictive distribution are derived from the stated probabilistic model; cited prior work is used only for standard kernel choices and as experimental baselines.

full rationale

The central derivations are self-contained. Equation (21) gives the factor posterior mean and covariance directly from the precision obtained by combining the likelihood (6) with the prior (8); Theorem 4 and Lemma 5 compute E_q[G^(d)G^(d)^T] from those same posterior parameters rather than from a fitted target; and the ELBO in (42) is assembled from the derived updates. The ARD hyperpriors (9)-(10) are standard sparsity-inducing priors, and the rank/feature pruning is an operational threshold on the inferred posterior variance, not a quantity fitted to the prediction target. The predictive Student-t distribution in Appendix 11 is an explicit approximation obtained by integrating the likelihood against q(W) and q(τ); whether its first-order variance is well calibrated is an accuracy question, not circularity. The synthetic experiment generates ground truth from the same model family, which is standard validation rather than circular reasoning. Self-citations (Batselier et al. 2017; Wesel and Batselier 2021, 2024) are used only to justify the polynomial/Gaussian kernel choices and to provide the T-KRR baseline; they do not supply the load-bearing derivation or forbid alternatives.

Assumptions & free parameters 8 free parameters · 4 assumptions · 0 invented entities

The method rests on the product-kernel structure and the mean-field approximation. No new entities are postulated. The free parameters are prior hyperparameters and initialization choices that are tuned per dataset.

free parameters (8)
  • Gamma hyperprior shape c0 = 1e-5 (most experiments)
    Controls sparsity of rank components; chosen by hand, adjusted per dataset.
  • Gamma hyperprior scale d0 = 1e-6 (most experiments)
    Controls sparsity of rank components; chosen by hand.
  • Gamma hyperprior shape g0 = 1e-6 (most experiments)
    Controls sparsity of feature dimensions; chosen by hand.
  • Gamma hyperprior scale h0 = 1e-6 (most experiments)
    Controls sparsity of feature dimensions; chosen by hand.
  • Noise precision prior a0, b0 = 1e-3 or 1e-2 depending on dataset
    Set to control overfitting; the paper says values can be adjusted per data set.
  • Initial rank R_max = 10, 25, or 50 depending on dataset
    User-specified initial capacity; affects performance.
  • Initial feature dimension M_max = 20, 25, 30, 40, or 50
    User-specified initial number of features per mode.
  • Truncation threshold = 1e-5 variance, after first 3 iterations
    Ad hoc criterion for pruning rank components.
assumptions (4)
  • domain assumption Feature map has Kronecker product form φ(x) = ⊗ φ^(d)(x^(d))
    Restricts model to product kernels; stated in Section 1, Eq (2).
  • domain assumption Mean-field factorization q(Θ) = ∏ q(W^(d)) q(λ_Md) q(λ_R) q(τ)
    Assumes independence across factor matrices and hyperparameters; standard VB assumption, stated in Eq (17).
  • standard math Gaussian priors with Gamma hyperpriors are conjugate
    Enables closed-form updates; standard result.
  • domain assumption The ARD prior induces the desired sparsity on CP columns and rows
    Empirically supported on synthetic data, but not guaranteed; relies on the Hadamard structure for columns.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Interpretable Bayesian Tensor Network Kernel Machines with Automatic Rank and Feature Selection." pith.science (2026). https://pith.science/paper/DWGCLGPH

@misc{pith2026250711136,
  author       = {Pith},
  title        = {Pith review of: Interpretable Bayesian Tensor Network Kernel Machines with Automatic Rank and Feature Selection},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DWGCLGPH}},
  note         = {Machine review of arXiv:2507.11136}
}
read the original abstract

Tensor Network (TN) Kernel Machines speed up model learning by representing parameters as low-rank TNs, reducing computation and memory use. However, most TN-based Kernel methods are deterministic and ignore parameter uncertainty. Further, they require manual tuning of model complexity hyperparameters like tensor rank and feature dimensions, often through trial-and-error or computationally costly methods like cross-validation. We propose Bayesian Tensor Network Kernel Machines, a fully probabilistic framework that uses sparsity-inducing hierarchical priors on TN factors to automatically infer model complexity. This enables automatic inference of tensor rank and feature dimensions, while also identifying the most relevant features for prediction, thereby enhancing model interpretability. All the model parameters and hyperparameters are treated as latent variables with corresponding priors. Given the Bayesian approach and latent variable dependencies, we apply a mean-field variational inference to approximate their posteriors. We show that applying a mean-field approximation to TN factors yields a Bayesian ALS algorithm with the same computational complexity as its deterministic counterpart, enabling uncertainty quantification at no extra computational cost. Experiments on synthetic and real-world datasets demonstrate the superior performance of our model in prediction accuracy, uncertainty quantification, interpretability, and scalability.

Figures

Figures reproduced from arXiv: 2507.11136 by the authors.

Figure 1
Figure 1. Representation of BTN-Kernel machines with the CPD-decomposed weight vector w as a proba￾bilistic graphical model showing the hierarchical sparsity inducing priors over the factor matrices {W(d) } D d=1 by the sparsity parameters λR and {λMd } D d=1. The dashed node denotes the observed data y, while the solid nodes represent random variables. Shape and scale hyperparameters of the Gamma priors placed on λR, {λMd } … view at source ↗
Figure 2
Figure 2. Representation of BTN-Kernel machines with the TT-decomposed weight vector w as a prob￾abilistic graphical model showing the hierarchical sparsity inducing priors over the core tensors {W(d) } D d=1 by the sparsity parameters {λR} D d=2 and {λMd } D d=1. The dashed node denotes the observed data y, while the solid nodes represent random variables. Shape and scale hyperpa￾rameters of the Gamma priors placed on {λR} D… view at source ↗
Figure 3
Figure 3. Posterior means of the factor matrices W˜ (d) for all d ∈ [1, 3], along with λ˜R and λ˜Md . The row labels to the left of each matrix show the posterior mean of λ˜Md , specific to each factor matrix. The column labels below the matrices represent the posterior mean of λ˜R, shared across all factor matrices. The model correctly identifies the tensor rank and feature dimensions by assigning larger precision values to … view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Reff and the variational lower bound (LB) during the training for varying initial rank values Rmax. The variational LB and Reff quickly stabilize early in training, with all models converging to low-rank solutions regardless of the initial Rmax. which by definition sho…
Figure 5
Figure 5. Figure 5: The average feature dimension, M¯ eff, for each feature across the data sets when ΛMd is on and off. The average feature dimension M¯ eff is consistently lower when the row-wise penalization term ΛMd is applied. RMSE NLL Reff Dataset Mmax λM-on λM-off λM-on λM-off λM-o…

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. (MPO)$^2$: Multivariate Polynomial Optimization based on Matrix Product Operators

    cs.LG 2026-07 conditional novelty 6.0 of 10

    Writing both feature maps and polynomial coefficient tensors as MPOs yields feature-order-independent polynomial models that beat CPD/TT polynomial baselines on most benchmarks.

  2. Tensor Network Kernel Machines: A JAX Framework for Machine Learning and Nonlinear System Identification

    cs.MS 2026-08 conditional novelty 5.0 of 10

    The tnkm JAX library provides a unified, modular implementation of tensor network kernel machines and reports competitive benchmark accuracy with low training cost.

Reference graph

Works this paper leans on

50 extracted references · 49 canonical work pages · cited by 2 Pith papers

  1. [1]

    Batselier, Z

    K. Batselier, Z. Chen, and N. Wong. Tensor Network alternating linear scheme for MIMO Volterra system identification. Automatica, 84: 0 26--35, 2017

    work page 2017

  2. [2]

    C. M. Bishop. Pattern Recognition and Machine Learning. Springer, New York, 2006. ISBN 978-0-387-31073-2

    work page 2006

  3. [3]

    Eckart - Young

    J. D. Carroll and J.-J. Chang. Analysis of individual differences in multidimensional scaling via an n-way generalization of “ Eckart - Young ” decomposition. Psychometrika, 35 0 (3): 0 283--319, 1970

    work page 1970

  4. [4]

    Cheng, Y

    L. Cheng, Y. C. Wu, and H. V. Poor. Probabilistic Tensor Canonical Polyadic Decomposition with Orthogonal Factors . IEEE Transactions on Signal Processing, 65 0 (3): 0 663--676, 2017

    work page 2017

  5. [5]

    Chu and Z

    W. Chu and Z. Ghahramani. Probabilistic Models for Incomplete Multi -dimensional Arrays . In Proceedings of the Twelfth International Conference on Artificial Intelligence and Statistics , pages 89--96. PMLR, 2009

    work page 2009

  6. [6]

    T. Dao, C. D. Sa, and C. Ré. Gaussian quadrature for kernel features. In Proceedings of the 31st International Conference on Neural Information Processing Systems (NIPS), pages 6109--6119, Red Hook, NY, USA, 2017. Curran Associates Inc

    work page 2017

  7. [7]

    Drineas and M

    P. Drineas and M. W. Mahoney. On the nyström method for approximating a gram matrix for improved kernel-based learning. Journal of Machine Learning Research, 6 0 (72): 0 2153--2175, 2005

    work page 2005

  8. [8]

    Garriga-Alonso, C

    A. Garriga-Alonso, C. E. Rasmussen, and L. Aitchison. Deep convolutional networks as shallow gaussian processes. In Proceedings of the International Conference on Learning Representations (ICLR), 2018

    work page 2018

Show all 50 references
  1. [9]

    Guhaniyogi, S

    R. Guhaniyogi, S. Qamar, and D. B. Dunson. Bayesian Tensor Regression . Journal of Machine Learning Research, 18 0 (79): 0 1--31, 2017

    2017

  2. [10]

    Hammer and K

    B. Hammer and K. Gersmann. A note on the universal approximation capability of support vector machines. Neural Processing Letters, 17 0 (1): 0 43--53, 2003

    2003

  3. [11]

    R. A. Harshman. Foundations of the parafac procedure: Model and conditions for an 'explanatory' multi-mode factor analysis. Technical Report 16, UCLA Working Papers in Phonetics, 1970

    1970

  4. [12]

    Hensman, N

    J. Hensman, N. Durrande, and A. Solin. Variational fourier features for gaussian processes. Journal of Machine Learning Research, 18 0 (1): 0 5537--5588, 2017

    2017

  5. [13]

    J. L. Hinrich and M. Morup. Probabilistic Tensor Train Decomposition . In 2019 27th European Signal Processing Conference ( EUSIPCO ) , pages 1--5, A Coruna, Spain, 2019. IEEE

    2019

  6. [14]

    J. L. Hinrich, K. H. Madsen, and M. Mørup. The probabilistic tensor decomposition toolbox. Machine Learning: Science and Technology, 1 0 (2), 2020

    2020

  7. [15]

    P. D. Hoff. Multilinear tensor regression for longitudinal relational data. The Annals of Applied Statistics, 9 0 (3), 2015

    2015

  8. [16]

    P. D. Hoff. Equivariant and Scale - Free Tucker Decomposition Models . Bayesian Analysis, 11 0 (3): 0 627--648, 2016

    2016

  9. [17]

    Izmailov, A

    P. Izmailov, A. Novikov, and D. Kropotov. Scalable gaussian processes with billions of inducing inputs via tensor train decomposition. In A. Storkey and F. Perez-Cruz, editors, Proceedings of the Twenty-First International Conference on Artificial Intelligence and Statistics, ...

    2018

  10. [18]

    P. J. H. Jørgensen, S. F. V. Nielsen, J. L. Hinrich, M. N. Schmidt, K. H. Madsen, and M. Mørup. Probabilistic PARAFAC2 , 2018

    2018

  11. [19]

    P. J. H. Jørgensen, S. F. V. Nielsen, J. L. Hinrich, M. N. Schmidt, K. H. Madsen, and M. Mørup. Analysis of Chromatographic Data using the Probabilistic PARAFAC2 . Proceedings of Second Workshop on Machine Learning and the Physical Sciences, 2019

    2019

  12. [20]

    S. A. Khan and S. Kaski. Bayesian multi-view tensor factorization. In T. Calders, editor, Proceedings of the Joint European Conference on Machine Learning and Knowledge Discovery in Databases (ECML PKDD), volume 8724 of Lecture Notes in Computer Science, pages 656--671. Springer, 2014

    2014

  13. [21]

    S. A. Khan, E. Leppäaho, and S. Kaski. Bayesian multi-tensor factorization. Machine Learning, 105 0 (2): 0 233--253, 2016

    2016

  14. [22]

    T. G. Kolda and B. W. Bader. Tensor Decompositions and Applications . SIAM Review, 51 0 (3): 0 455--500, 2009

    2009

  15. [23]

    Konstantinidis, Y

    K. Konstantinidis, Y. Xu, Q. Zhao, and D. Mandic. Variational bayesian tensor networks with structured posteriors. In Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pages 3638--3642, 2022

    2022

  16. [24]

    J. B. Kruskal. Three-way arrays: Rank and uniqueness of trilinear decompositions, with application to arithmetic complexity and statistics. Linear Algebra and Its Applications, 18 0 (2): 0 95--138, 1977

    1977

  17. [25]

    J. W. Lee, Y. Bahri, R. Novak, S. S. Schoenholz, J. Pennington, and J. Sohl-Dickstein. Deep neural networks as gaussian processes. arXiv preprint arXiv:1711.00165, 2018. cs, stat

    2018 arXiv

  18. [26]

    D. J. C. MacKay. Bayesian methods for backpropagation networks. In E. Domany, J. L. van Hemmen, and K. Schulten, editors, Models of Neural Networks III, chapter 6, pages 211--254. Springer, 1994

    1994

  19. [27]

    Mutn \'y and A

    M. Mutn \'y and A. Krause. Efficient high dimensional bayesian optimization with additivity and quadrature fourier features. In Proceedings of the 32nd International Conference on Neural Information Processing Systems (NIPS), pages 9019--9030. Curran Associates Inc., 2018

    2018

  20. [28]

    R. M. Neal. Bayesian Learning for Neural Networks, volume 118 of Lecture Notes in Statistics. Springer New York, New York, NY, 1996

    1996

  21. [29]

    Novak, L

    R. Novak, L. Xiao, Y. Bahri, J. Lee, G. Yang, J. Hron, D. A. Abolafia, J. Pennington, and J. Sohl-Dickstein. Bayesian deep convolutional networks with many channels are gaussian processes. In Proceedings of the International Conference on Learning Representations (ICLR), 2018

    2018

  22. [30]

    Novikov, M

    A. Novikov, M. Trofimov, and I. Oseledets. Exponential machines. Bulletin of the Polish Academy of Sciences Technical Sciences, 66 0 (No 6 (Special Section on Deep Learning: Theory and Practice)): 0 789--797, 2018

    2018

  23. [31]

    I. V. Oseledets. Tensor-train decomposition. SIAM Journal on Scientific Computing, 33 0 (5): 0 2295--2317, 2011

    2011

  24. [32]

    Rahimi and B

    A. Rahimi and B. Recht. Random features for large-scale kernel machines. In Proceedings of the 20th International Conference on Neural Information Processing Systems (NIPS), pages 1177--1184, Red Hook, NY, USA, 2007. Curran Associates Inc

    2007

  25. [33]

    P. Rai, Y. Wang, S. Guo, G. Chen, D. Dunson, and L. Carin. Scalable Bayesian Low - Rank Decomposition of Incomplete Multiway Tensors . In Proceedings of the 31st International Conference on Machine Learning , pages 1800--1808. PMLR, 2014

    2014

  26. [34]

    P. Rai, Y. Wang, and L. Carin. Leveraging Features and Networks for Probabilistic Tensor Decomposition . Proceedings of the AAAI Conference on Artificial Intelligence, 29 0 (1), 2015

    2015

  27. [35]

    C. E. Rasmussen and C. K. I. Williams. Gaussian Processes for Machine Learning. Adaptive Computation and Machine Learning. MIT Press, Cambridge, MA, 2006

    2006

  28. [36]

    Solin and S

    A. Solin and S. S \"a rkk \"a . Hilbert space methods for reduced-rank gaussian process regression. Statistics and Computing, 30 0 (2): 0 419--446, 2020

    2020

  29. [37]

    J. A. K. Suykens and J. Vandewalle. Least squares support vector machine classifiers. Neural Processing Letters, 9 0 (3): 0 293--300, 1999

    1999

  30. [38]

    J. A. K. Suykens, T. V. Gestel, J. D. Brabanter, B. D. Moor, and J. Vandewalle. Least Squares Support Vector Machines. World Scientific, Singapore, 2002

    2002

  31. [39]

    L. R. Tucker. Some mathematical notes on three-mode factor analysis. Psychometrika, 31 0 (3): 0 279--311, 1966

    1966

  32. [40]

    Wesel and K

    F. Wesel and K. Batselier. Large-scale learning with fourier features and tensor decompositions. In Proceedings of the 35th International Conference on Neural Information Processing Systems , pages 17543--17554, 2021

    2021

  33. [41]

    Wesel and K

    F. Wesel and K. Batselier. Quantized fourier and polynomial features for more expressive tensor network models, 2024

    2024

  34. [42]

    C. K. I. Williams and M. Seeger. Using the nyström method to speed up kernel machines. In Proceedings of the 14th International Conference on Neural Information Processing Systems (NIPS), pages 682--688. MIT Press, 2001

    2001

  35. [43]

    J. M. Winn and C. M. Bishop. Variational message passing. Journal of Machine Learning Research, 6: 0 661--694, 2005

    2005

  36. [44]

    Xiong, X

    L. Xiong, X. Chen, T.-K. Huang, J. Schneider, and J. G. Carbonell. Temporal Collaborative Filtering with Bayesian Probabilistic Tensor Factorization . In Proceedings of the 2010 SIAM International Conference on Data Mining , pages 211--222. Society for Industrial and Applied M...

    2010

  37. [45]

    Z. Xu, F. Yan, and Y. Qi. Bayesian Nonparametric Models for Multiway Data Analysis . IEEE Trans. Pattern Anal. Mach. Intell., 37 0 (2): 0 475--487, 2015

    2015

  38. [46]

    Yang and D

    Y. Yang and D. B. Dunson. Bayesian Conditional Tensor Factorizations for High - Dimensional Classification . Journal of the American Statistical Association, 111 0 (514): 0 656--669, 2016

    2016

  39. [47]

    R. Yu, G. Li, and Y. Liu. Tensor Regression Meets Gaussian Processes . In Proceedings of the Twenty-First International Conference on Artificial Intelligence and Statistics, pages 482--490. PMLR, 2018

    2018

  40. [48]

    Q. Zhao, L. Zhang, and A. Cichocki. Bayesian CP factorization of incomplete tensors with automatic rank determination. IEEE Transactions on Pattern Analysis and Machine Intelligence, 37 0 (9): 0 1751--1763, 2015 a

    2015

  41. [49]

    Q. Zhao, L. Zhang, and A. Cichocki. Bayesian Sparse Tucker Models for Dimension Reduction and Tensor Completion , 2015 b

    2015

  42. [50]

    Q. Zhao, G. Zhou, L. Zhang, A. Cichocki, and S. ichi Amari. Bayesian Robust Tensor Factorization for Incomplete Multiway Data . IEEE Transactions on Neural Networks and Learning Systems, 27 0 (4): 0 736--748, 2016

    2016

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.