Pith. sign in

REVIEW 4 major objections 3 minor 50 references

Integrating Random Effects in Variational Autoencoders for Dimensionality Reduction of Correlated Data

T0 review · 4 major / 3 minor · reviewed 2026-08-11 · deepseek-v4-flash

Pith's one-line read Adding random effects to variational autoencoders makes them far better at reconstructing correlated data, the paper reports.

desk verdict LMMVAE is a genuinely useful idea with strong empirical results for categorical and longitudinal data, but the spatial variant's loss is not a valid ELBO under the correlated prior and should be re-derived or qualified. read the letter →

arxiv 2412.16899 v2 pith:2T3QWL4G submitted 2024-12-22 stat.ML cs.LG

classification stat.MLcs.LG
keywords variationalautoencoderrandomeffectslinearmixedmodelscorrelateddatadimensionalityreductionreconstructionerrorlatentrepresentationslongitudinalandspatial
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

This paper argues that the standard variational autoencoder's assumption of independent observations is the reason it underperforms on datasets with clustering, temporal, or spatial structure. It proposes LMMVAE, a variational autoencoder whose decoder output is $f(U)+ZB$, where $U$ is the usual independent latent variable and $B$ is a matrix of random effects shared by observations in the same group, in the spirit of linear mixed models. On held-out data from simulations and real tabular datasets, the model reports substantially lower squared reconstruction error and negative log-likelihood than standard VAEs, PCA baselines, and Gaussian-process-prior VAE variants, and the learned representations improve a downstream classification task. If the paper is right, a small architectural change—adding a random-effects term with its own encoder—is a general way to make deep generative dimensionality reduction respect known correlation structure.

What carries the argument

The load-bearing object is the matrix model $X = f(U) + ZB + E$, with $B \sim \mathrm{MN}(0, K, D)$. Here $Z$ is a user-specified $n\times q$ design matrix: an indicator matrix for categorical levels, polynomial-in-time columns for longitudinal data, and location indicators for spatial data; $D$ is a diagonal $p\times p$ prior variance matrix for the $p$ feature-wise random effects; and $K$ is a $q\times q$ between-group correlation kernel, the identity matrix for categorical and longitudinal settings and an RBF kernel for spatial data. The architecture uses an encoder that outputs the mean and variance of $B$, samples $B$ by the reparameterization trick, averages the samples per group to form $B$ for the mini-batch, and adds $ZB$ to the fixed decoder output. The loss in equation (9) is the ELBO with the KL term decomposed into fixed and random parts, so the only change relative to a standard VAE is one extra KL term per random-effect source. For spatial data, the paper enforces between-row correlation by left-multiplying the sampled $B$ by the Cholesky factor of the posterior covariance $\Psi_*$ obtained from the best linear unbiased predictor formula.

What would settle it

Compute the exact KL divergence between the correlated prior $B \sim \mathrm{MN}(0, K, D)$ and the factorized variational posterior used by the spatial model, and compare the resulting ELBO with equation (9); if the extra terms involving $K$ and $\Psi_*$ are non-negligible, the reported spatial gains are not coming from maximizing the stated lower bound.

Watch

Extended reading notes

Core claim

The central claim is that the generative model $X = f(U) + ZB + E$, where $U$ are ordinary independent latent variables and $B$ is a matrix of random effects with a matrix-normal prior $\mathrm{MN}(0, K, D)$, is a strictly better VAE-type model for correlated data. The fixed part $f(U)$ is the usual non-linear decoder; the design matrix $Z$ encodes which cluster, subject, location, or time polynomial each row belongs to, and $B$ captures the shared deviations for those groups. The paper derives a modified ELBO whose KL divergence splits into a fixed and a random part under the assumption that $q_\phi(u,b\mid x)=q_\phi(u\mid x)q_\phi(b\mid x)$, yielding the loss in equation (9), and implements the model with separate encoders for $u$ and $b$. Across simulated high-cardinality categorical, longitudinal, and spatial settings, and on real datasets such as news shares, store sales, census income, and used-car listings, LMMVAE reports the best test reconstruction error and NLL, often by a wide margin, and better $k$-NN classification on the latent space for small latent dimension. On CelebA images with face identity as the random effect, the method matches but does not beat a VAE that ignores the identity, a result the paper attributes to the need for additional external features.

Load-bearing premise

For spatial data, the paper assumes that the loss derived for independent random effects remains the correct variational objective when the random effects are instead correlated across locations through a kernel matrix; if that loss is not the true lower bound, the spatial experiments train a heuristic rather than the stated model.

Editorial extensions

If this is right

  • On datasets with known grouping structure, practitioners can expect LMMVAE to reconstruct held-out rows markedly better than a VAE that ignores the grouping, because the random-effect term carries the group-level shared signal.
  • The learned random-effects matrix $\hat{B}$ can be read as per-group offsets for each feature, giving a built-in diagnostic such as mapping location effects onto a map without fitting a separate model.
  • The same architecture covers categorical, longitudinal, spatial, and combined scenarios by changing only $Z$ and the prior on $B$, so one implementation handles several correlation types.
  • For downstream tasks, the fixed latent representation $U$ from LMMVAE supports better supervised classification at small latent dimension than standard VAE or PCA representations on the tested datasets.
  • On image data with a categorical random effect, the paper's experiments show negligible gains over ignoring the effect, suggesting the benefit concentrates on tabular data or requires additional external features.

Reading between the lines

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

  • If the spatial loss in equation (9) is not the exact ELBO for the correlated prior $B \sim \mathrm{MN}(0, K, D)$, the spatial results may still stand as a heuristic, but the claim of maximizing a variational lower bound in that setting would need re-derivation; a direct computation of the KL divergence would settle it.
  • The additive random-effects decomposition could be transferred to other deep generative models, such as normalizing flows or diffusion decoders, whenever rows are exchangeable within known groups, because the argument relies only on the additive structure in the decoder.
  • Treating $D$ and $K$ as learnable variance components rather than fixed priors might improve performance further and would connect LMMVAE to empirical-Bayes mixed-model estimation; the paper fixes $D=\sigma_b^2 I$ and the RBF length scale.
  • The independence assumption between $U$ and $B$ given $X$ could be relaxed by sharing encoder features, and the paper's own observation that a single double-output encoder is slightly inferior suggests the separation itself, not just the architecture, is doing the work.
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

4 major / 3 minor

Summary. The paper proposes LMMVAE, a variational autoencoder that augments the standard fixed latent model with a linear mixed-model random-effect term, so that the decoder generates x = f(u) + Zb + ε. The authors modify the ELBO to include separate encoders for the fixed latent u and the random effect b, and they derive an explicit loss for the high-cardinality categorical case with an independent random-effect prior B ~ MN(0, I_q, D), D diagonal. They then extend this construction by analogy to longitudinal data (multiple polynomial random effects) and spatial data (B ~ MN(0, K, D) with an RBF kernel), reporting substantial improvements in squared reconstruction error and negative ELBO loss on unseen data across simulated and real tabular datasets, plus a CelebA image experiment and downstream classification gains.

Significance. The idea of separating fixed and random latent effects is a natural and practically motivated extension of VAEs to correlated tabular data, and the empirical evaluation is unusually broad: multiple correlation scenarios, six real tabular datasets, comparisons to PCA, VAE variants, VRAE, and SGP-VAE, and publicly available code. The categorical and longitudinal derivations in Appendices A and B are standard and appear correct. However, the spatial variant, one of the three headline scenarios, is not derived as an ELBO under the correlated prior, and the mini-batch averaging step is a heuristic. These gaps weaken the theoretical grounding of a substantial portion of the empirical claims and need to be addressed before the paper can be accepted.

major comments (4)
  1. [Section 3.3.3, Eq. (9)] The spatial variant changes the random-effect prior to B ~ MN(0, K, D) with an RBF kernel K but states, without derivation, that 'the loss function, therefore, remains the same as in (9)'. Equation (9) is derived in Appendix B only for B ~ MN(0, I_q, D) with D diagonal. For B ~ MN(0, K, D), the KL divergence between a diagonal Gaussian q(b|x) and N(0, K) contains terms involving K^{-1}, log|K|, and Tr(K^{-1}Σ), none of which appear in Eq. (9). As written, the spatial objective is not the ELBO of the spatial generative model, so the spatial experiments in Tables 1, 13–16, and 23–25 train a heuristic rather than a variational lower bound. The paper must either derive the spatial ELBO correctly or explicitly reframe the spatial method as an approximation with its own objective.
  2. [Section 3.1, architecture description] The model samples one p-dimensional b vector per observation in a mini-batch, but then averages these vectors by group level to form the B matrix before computing ZB. This averaging is a heuristic that does not follow from the ELBO in Eq. (8), where each observation x_ij has its own b_j in the generative model. The paper does not show that this averaging step is equivalent to optimizing a coherent variational objective, either for the categorical/longitudinal cases or for the spatial case. Since the group-level B is used both in training and in the prediction procedure, a derivation or a clear statement that this is an approximation is needed.
  3. [Section 3.3.3, 'Unknown mode' experiments] For unseen locations, the paper uses the Cholesky factor Ψ*^{1/2} computed from training locations and left-multiplies the sampled B. This is a kriging-style heuristic: no variational posterior for new random effects is defined, and the loss for test observations is not the ELBO conditional on the training posterior. The Unknown-mode results in Tables 14 and 23 should either be accompanied by a derivation or explicitly labeled as a predictive heuristic, so that the reader can interpret the reported gains appropriately.
  4. [Section 3.1, Eq. (6) and abstract] The claim that the matrix formulation in Eq. (6) is 'general enough' to cover categorical, longitudinal, and spatial scenarios is stronger than what is actually derived. The loss derivations cover the independent-prior categorical case (Appendix B); the longitudinal extension changes the prior to MN(0, Φ ⊗ I_q, D) without computing the corresponding KL, and the spatial case changes the row covariance to K without re-deriving the KL. Please state precisely which scenarios have a fully derived ELBO and which are heuristic extensions, and adjust the abstract and Section 3.1 accordingly.
minor comments (3)
  1. [Throughout] The paper uses 'negative log-likelihood (NLL)' to refer to the negative ELBO, but the reported quantities are bounds on the marginal likelihood, not the marginal likelihood itself. Please use 'negative ELBO' consistently or define the terminology explicitly.
  2. [Table 2] The column header 'n_j' is confusing for datasets with multiple categorical features (e.g., News and Spotify), where 'q' is reported as two or more values and 'n_j' is not a single number. Consider clarifying the column meaning or splitting it into group-size descriptions per feature.
  3. [Section 3.3.3] The symbol Ψ*^{1/2} is used without specifying the Cholesky convention (lower vs. upper triangular). For reproducibility, please state the convention used in the implementation.

Circularity Check

0 steps flagged · score 1.0 of 10

No circular derivation: LMMVAE loss is derived from the ELBO and evaluated out-of-sample; the spatial loss assertion is an omitted-proof correctness issue, not circularity.

full rationale

The derivation chain in Sections 3.1-3.2 and Appendices A-B is self-contained: Eq (8) follows from the factorized surrogate q(u,b|x)=q(u|x)q(b|x), and Eq (9) is obtained by evaluating the KL between two diagonal Gaussians for the independent categorical prior. The central empirical claims are out-of-sample comparisons: models are trained on (Xtr,Ztr) and scored on held-out (Xte,Zte), so the reported reconstruction-error and NLL improvements are not identities forced by the loss definition. The simulated data are generated from the same LMM family as the model, which is a recoverability sanity check rather than a circular step. The paper cites the authors' prior LMM-DNN work, but those citations are not load-bearing: the ELBO decomposition and loss derivation do not appeal to them, and no uniqueness theorem is imported from them. One passage deserves explicit flagging: Section 3.3.3 changes the RE prior to MN(0,K,D) and states 'The loss function, therefore, remains the same as in (9)' without deriving the KL for the correlated prior; this is a genuine omitted derivation and a correctness risk for the spatial experiments, but it is not circularity because Eq (9) is not defined in terms of the spatial prediction target. Overall, no prediction or first-principles result reduces to its own input, so the circularity score is minimal.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The model relies on standard ELBO machinery plus several modeling choices. The main contributed structure is the random-effect latent variable b with a matrix-normal prior. Free parameters include the KL weight β, the prior variance σ_b^2, and the kernel length-scale l^2. The spatial variant and the mini-batch averaging are not fully derived from the ELBO.

free parameters (4)
  • β (KL weight) = 0.01 in all reported experiments; tuned for real data
    Scales the KL divergence terms in the loss; chosen by validation, not derived.
  • σ_b^2 prior (random-effect prior variance) = 1.0 in simulations; 0.001 to 1.0 across real datasets (Table 2)
    Prior variance for random effects; set to 1 or tuned per dataset, affects the KL term.
  • Kernel length-scale l^2 in spatial model = 1.0 (assumed known)
    RBF kernel length-scale; fixed to 1 rather than estimated, but true values in simulations differ (0.3 or 3.0).
  • n_samp for Ψ* approximation = 10000
    Number of rows sampled to compute the V matrix inverse for the spatial BLUP; a computational approximation.
assumptions (5)
  • domain assumption Fixed and random latent variables are conditionally independent given data: q(u,b|x)=q(u|x)q(b|x)
    Section 3.1 (Equation 7); enables loss decomposition into separate KL terms.
  • domain assumption Random effects B follow a matrix normal prior with known covariance structure
    Section 3.1 (Equation 6); the prior D is set to I in training, but the generative model may differ.
  • domain assumption The generative model X = f(U) + ZB + E holds exactly
    Section 3.1 (Equation 6); simulations use this model, real data may violate it.
  • ad hoc to paper Spatial B posterior can be approximated by the LMM BLUP with known kernel
    Section 3.3.3; uses Ψ* from LMM without deriving the corresponding ELBO.
  • ad hoc to paper Mini-batch averaging of per-observation b samples yields the group-level random effect
    Section 3.1, operation ⊙; not derived from the ELBO, but used in the forward pass.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Integrating Random Effects in Variational Autoencoders for Dimensionality Reduction of Correlated Data." pith.science (2026). https://pith.science/paper/2T3QWL4G

@misc{pith2026241216899,
  author       = {Pith},
  title        = {Pith review of: Integrating Random Effects in Variational Autoencoders for Dimensionality Reduction of Correlated Data},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2T3QWL4G}},
  note         = {Machine review of arXiv:2412.16899}
}
read the original abstract

Variational Autoencoders (VAE) are widely used for dimensionality reduction of large-scale tabular and image datasets, under the assumption of independence between data observations. In practice, however, datasets are often correlated, with typical sources of correlation including spatial, temporal and clustering structures. Inspired by the literature on linear mixed models (LMM), we propose LMMVAE -- a novel model which separates the classic VAE latent model into fixed and random parts. While the fixed part assumes the latent variables are independent as usual, the random part consists of latent variables which are correlated between similar clusters in the data such as nearby locations or successive measurements. The classic VAE architecture and loss are modified accordingly. LMMVAE is shown to improve squared reconstruction error and negative likelihood loss significantly on unseen data, with simulated as well as real datasets from various applications and correlation scenarios. It also shows improvement in the performance of downstream tasks such as supervised classification on the learned representations.

Figures

Figures reproduced from arXiv: 2412.16899 by the authors.

Figure 1
Figure 1. LMMVAE architecture: data X enters two separate FE and RE encoders (or a single encoder with double output), to produce the fixed LV u and RE b by the reparameterization trick. u goes through the FE decoder, its output f(u) is added the RE term ZB after Z enters the model and multiplies the RE matrix B after it had been properly formed from the b RE vectors, as depicted by the ⊙ symbol (see the different covariance … view at source ↗
Figure 2
Figure 2. Predicted vs. true scatter plots for simulated datasets with [PITH_FULL_IMAGE:figures/full_fig_p009_2.png] view at source ↗
Figure 3
Figure 3. Exploring the Bˆ RE matrix from the Cars dataset containing spatial features, with a raster plot. Left: distribution across the US of the Bˆ column corresponding to the price feature; Right: distribution across the US of the Bˆ column corresponding to the odometer feature. 4.2 Real Data We used two real datasets for each covariance scenario, their full details can be found in Appendix C, and some technical details c… view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Comparing true vs. reconstructed Xte for the Rossmann stores longitudinal dataset, Future mode with d = 1. In this mode the model is trained on the first 25 months of dataset to reconstruct the last 6 months of the dataset. Left: comparing LMMVAE and VAE with entity em…
Figure 5
Figure 5. Figure 5: Comparing true vs. reconstructed Xte for the CelebA dataset of facial images with d = 100, using convolutional neural networks for both FE and RE encoders. Left: true faces; Right: reconstructed faces. 19 [PITH_FULL_IMAGE:figures/full_fig_p019_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

50 extracted references · 31 canonical work pages

  1. [1]

    Martín Abadi, Ashish Agarwal, Paul Barham, Eugene Brevdo, Zhifeng Chen, Craig Citro, Greg S. Corrado, Andy Davis, Jeffrey Dean, Matthieu Devin, Sanjay Ghemawat, Ian Goodfellow, Andrew Harp, Geoffrey Irving, Michael Isard, Yangqing Jia, Rafal Jozefowicz, Lukasz Kaiser, Manjunath Kudlur, Josh Levenberg, Dandelion Mané, Rajat Monga, Sherry Moore, Derek Murra...

  2. [2]

    Variational autoencoders: A hands-off approach to volatility

    Maxime Bergeron, Nicholas Fung, John Hull, Zissis Poulos, and Andreas Veneris. Variational autoencoders: A hands-off approach to volatility. The Journal of Financial Data Science, 4 (2):125–138, 2022. ISSN 2640-3943. doi: 10.3905/jfds.2022.1.093. URL https://jfds. pm-research.com/content/4/2/125

  3. [3]

    Bowman, Luke Vilnis, Oriol Vinyals, Andrew M

    Samuel R. Bowman, Luke Vilnis, Oriol Vinyals, Andrew M. Dai, Rafal Jozefowicz, and Samy Bengio. Generating sentences from a continuous space, 2015. URL https://arxiv.org/ abs/1511.06349

  4. [4]

    Gaussian process prior variational autoencoders

    Francesco Paolo Casale, Adrian Dalca, Luca Saglietti, Jennifer Listgarten, and Nicolo Fusi. Gaussian process prior variational autoencoders. In S. Bengio, H. Wallach, H. Larochelle, K. Grauman, N. Cesa-Bianchi, and R. Garnett, editors, Advances in Neural Information Pro- cessing Systems, volume 31. Curran Associates, Inc., 2018. URL https://proceedings. n...

  5. [5]

    National environmental public health tracking network data explorer - asthma in adults, Nov 2017

    CDC. National environmental public health tracking network data explorer - asthma in adults, Nov 2017. URL https://www.cdc.gov/nceh/tracking/topics/asthma.htm

  6. [6]

    François Chollet et al. Keras. https://keras.io, 2015

  7. [7]

    Noel A. C. Cressie. Statistics for spatial data. Wiley series in probability and statistics. Wiley- Interscience Publication, New York, revised edition.. edition, 1993. ISBN 1-119-11515-9

  8. [8]

    Converse, Jeff Hajewski, and Suely Oliveira

    Mariana Curi, Geoffrey A. Converse, Jeff Hajewski, and Suely Oliveira. Interpretable variational autoencoders for cognitive models. In 2019 International Joint Conference on Neural Networks (IJCNN), pages 1–8, 2019. doi: 10.1109/IJCNN.2019.8852333

Show all 50 references
  1. [9]

    Tutorial on variational autoencoders, 2016

    Carl Doersch. Tutorial on variational autoencoders, 2016. URL https://arxiv.org/abs/ 1606.05908

  2. [10]

    Variational recurrent auto-encoders

    Otto Fabius and Joost R Van Amersfoort. Variational recurrent auto-encoders. arXiv preprint arXiv:1412.6581, 2014

  3. [11]

    Matrix variate distributions

    Arjun K Gupta and Daya K Nagar. Matrix variate distributions. Chapman and Hall/CRC, 2018

  4. [12]

    Hancock and Taghi M

    John T. Hancock and Taghi M. Khoshgoftaar. Survey on categorical data for neural networks. Journal of Big Data, 7(1):28, Apr 2020. ISSN 2196-1115. doi: 10.1186/s40537-020-00305-w. URL https://doi.org/10.1186/s40537-020-00305-w

  5. [13]

    beta-V AE: Learning basic visual concepts with a constrained variational framework

    Irina Higgins, Loic Matthey, Arka Pal, Christopher Burgess, Xavier Glorot, Matthew Botvinick, Shakir Mohamed, and Alexander Lerchner. beta-V AE: Learning basic visual concepts with a constrained variational framework. In International Conference on Learning Representations,

  6. [14]

    Reducing the dimensionality of data with neural networks

    Geoffrey E Hinton and Ruslan R Salakhutdinov. Reducing the dimensionality of data with neural networks. science, 313(5786):504–507, 2006

  7. [15]

    Scalable gaussian process variational autoencoders

    Metod Jazbec, Matt Ashman, Vincent Fortuin, Michael Pearce, Stephan Mandt, and Gun- nar Rätsch. Scalable gaussian process variational autoencoders. In Arindam Banerjee and Kenji Fukumizu, editors, Proceedings of The 24th International Conference on Artificial Intel- ligence an...

  8. [16]

    Jolliffe

    I.T. Jolliffe. Principal Component Analysis. Springer Series in Statistics. Springer, 2002. ISBN 9780387954424

  9. [17]

    Airbnb price prediction using machine learning and sentiment analysis, 2019

    Pouya Rezazadeh Kalehbasti, Liubov Nikolenko, and Hoormazd Rezaei. Airbnb price prediction using machine learning and sentiment analysis, 2019

  10. [18]

    Auto-encoding variational bayes, 2013

    Diederik P Kingma and Max Welling. Auto-encoding variational bayes, 2013. URL https: //arxiv.org/abs/1312.6114

  11. [19]

    Kingma and Max Welling

    Diederik P. Kingma and Max Welling. An introduction to variational autoencoders.Foundations and Trends® in Machine Learning, 12(4):307–392, 2019. ISSN 1935-8237. doi: 10.1561/ 2200000056. URL http://dx.doi.org/10.1561/2200000056

  12. [20]

    Probabilistic non-linear principal component analysis with gaussian process latent variable models

    Neil Lawrence. Probabilistic non-linear principal component analysis with gaussian process latent variable models. Journal of Machine Learning Research, 6(60):1783–1816, 2005. URL http://jmlr.org/papers/v6/lawrence05a.html

  13. [21]

    Deep learning face attributes in the wild

    Ziwei Liu, Ping Luo, Xiaogang Wang, and Xiaoou Tang. Deep learning face attributes in the wild. In Proceedings of International Conference on Computer Vision (ICCV), December 2015

  14. [22]

    Auxiliary deep generative models

    Lars Maaløe, Casper Kaae Sønderby, Søren Kaae Sønderby, and Ole Winther. Auxiliary deep generative models. In Maria Florina Balcan and Kilian Q. Weinberger, editors, Proceedings of The 33rd International Conference on Machine Learning, volume 48 of Proceedings of Machine Learn...

  15. [23]

    McCulloch, Shayle R

    Charles E. McCulloch, Shayle R. Searle, and John M. Neuhaus. Generalized, Linear, and Mixed Models. John Wiley and Sons, Inc., June 2008. ISBN 978-0-470-07371-1

  16. [24]

    Tidy tuesday: A weekly data project aimed at the r ecosystem, 2022

    Thomas Mock. Tidy tuesday: A weekly data project aimed at the r ecosystem, 2022. URL https://github.com/rfordatascience/tidytuesday

  17. [25]

    Multi-source social feedback of online news feeds

    Nuno Moniz and Luis Torgo. Multi-source social feedback of online news feeds. CoRR, 2018

  18. [26]

    Us census demographic data, Mar 2019

    MuonNeutrino. Us census demographic data, Mar 2019. URL https://www.kaggle.com/ datasets/muonneutrino/us-census-demographic-data

  19. [27]

    Deep variational autoencoder for modeling functional brain networks and adhd identification

    Ning Qiang, Qinglin Dong, Yifei Sun, Bao Ge, and Tianming Liu. Deep variational autoencoder for modeling functional brain networks and adhd identification. In2020 IEEE 17th International Symposium on Biomedical Imaging (ISBI) , pages 554–557, 2020. doi: 10.1109/ISBI45749. 2020.9098480

  20. [28]

    A unifying view of sparse approxi- mate gaussian process regression

    Joaquin Quiñonero-Candela and Carl Edward Rasmussen. A unifying view of sparse approxi- mate gaussian process regression. Journal of Machine Learning Research, 6(65):1939–1959,

  21. [29]

    J. O. Ramsay and B.W. Silverman. Functional Data Analysis. Springer Series in Statistics. Springer, 2005. ISBN 9780387400808

  22. [30]

    Hierarchical variational models

    Rajesh Ranganath, Dustin Tran, and David Blei. Hierarchical variational models. In Maria Florina Balcan and Kilian Q. Weinberger, editors, Proceedings of The 33rd Inter- national Conference on Machine Learning , volume 48 of Proceedings of Machine Learn- ing Research, pages 32...

  23. [31]

    Carl Edward Rasmussen and Christopher K. I. Williams. Gaussian Processes for Machine Learning (Adaptive Computation and Machine Learning) . The MIT Press, 2005. ISBN 026218253X

  24. [32]

    Used cars dataset - vehicles listings from craigslist.org, 2020

    Austin Reese. Used cars dataset - vehicles listings from craigslist.org, 2020. URL https: //www.kaggle.com/datasets/austinreese/craigslist-carstrucks-data

  25. [33]

    Variational autoencoders pursue pca directions (by accident), 2018

    Michal Rolinek, Dominik Zietlow, and Georg Martius. Variational autoencoders pursue pca directions (by accident), 2018. URL https://arxiv.org/abs/1812.06775. 15

  26. [34]

    Rossmann store sales, 2016

    Rossmann. Rossmann store sales, 2016. URL https://www.kaggle.com/competitions/ rossmann-store-sales/

  27. [35]

    Nonlinear component analysis as a kernel eigenvalue problem

    Bernhard Schölkopf, Alexander Smola, and Klaus-Robert Müller. Nonlinear component analysis as a kernel eigenvalue problem. Neural computation, 10(5):1299–1319, 1998

  28. [36]

    Variance components

    Shayle R Searle, George Casella, and Charles McCulloch. Variance components. Wiley Series in Probability and Statistics. John Wiley & Sons, 1992

  29. [37]

    Using random effects to account for high-cardinality categorical features and repeated measures in deep neural networks

    Giora Simchoni and Saharon Rosset. Using random effects to account for high-cardinality categorical features and repeated measures in deep neural networks. In M. Ranzato, A. Beygelzimer, Y . Dauphin, P.S. Liang, and J. Wortman Vaughan, editors, Advances in Neural Information P...

  30. [38]

    Integrating random effects in deep neural networks

    Giora Simchoni and Saharon Rosset. Integrating random effects in deep neural networks. Journal of Machine Learning Research , 2023. doi: 10.48550/ARXIV .2206.03314. URL https://arxiv.org/abs/2206.03314. to appear

  31. [39]

    Uk biobank: An open access resource for identifying the causes of a wide range of complex diseases of middle and old age

    Cathie Sudlow, John Gallacher, Naomi Allen, Valerie Beral, Paul Burton, John Danesh, Paul Downey, Paul Elliott, Jane Green, Martin Landray, Bette Liu, Paul Matthews, Giok Ong, Jill Pell, Alan Silman, Alan Young, Tim Sprosen, Tim Peakman, and Rory Collins. Uk biobank: An open a...

  32. [40]

    Tipping and Christopher M

    Michael E. Tipping and Christopher M. Bishop. Probabilistic principal component analysis. Journal of the Royal Statistical Society. Series B (Statistical Methodology) , 61(3):611–622,

  33. [41]

    Bayesian deep net glm and glmm

    Minh-Ngoc Tran, Nghia Nguyen, David Nott, and Robert Kohn. Bayesian deep net glm and glmm. Journal of Computational and Graphical Statistics, 29(1):97–113, 2020. doi: 10.1080/ 10618600.2019.1637747. URL https://doi.org/10.1080/10618600.2019.1637747

  34. [42]

    Recent advances in variational autoencoders with represen- tation learning for biomedical informatics: A survey

    Ruoqi Wei and Ausif Mahmood. Recent advances in variational autoencoders with represen- tation learning for biomedical informatics: A survey. IEEE Access, 9:4939–4956, 2021. doi: 10.1109/ACCESS.2020.3048309

  35. [43]

    Kim, and Vikas Singh

    Yunyang Xiong, Hyunwoo J. Kim, and Vikas Singh. Mixed effects neural networks (menets) with applications to gaze estimation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), June 2019. 16 A Proof of KL-divergence decomposition Mark z...

  36. [47]

    High-cardinality categorical features include artist ( q = 10K), al- bum (q = 22K), playlist (q = 2.3K) and subgenre (q = 553)

    28K songs with 12 audio features. High-cardinality categorical features include artist ( q = 10K), al- bum (q = 22K), playlist (q = 2.3K) and subgenre (q = 553). Longitudinal Rossmann Kaggle (Free) [34] Total monthly sales in $ from over q = 1.1K stores around Europe, over 25-...

  37. [48]

    Time-varying features include gender, age, height, different food intakes, smoking habits and many more

    q = 469K subjects of the UK Biobank cohort for which we have 1-4 systolic blood pressure (SBP) measures. Time-varying features include gender, age, height, different food intakes, smoking habits and many more. Spatial Income Kaggle (Free) [26] Mean yearly income in $ for 71K U...

  38. [49]

    50K Airbnb listings in NYC scraped by [17], ETL follows their steps exactly. In addition to longitude and latitude (overall q = 3K unique locations), features include floor number, neighborhood, some top 1-ngram tokens counts from description and more, as well as the listing’s...

  39. [50]

    wears glasses

    202K facial images from 10,177 celebrities (q = 10K). Additional features not considered here are binary features such as “wears glasses” and continuous features such as pixel nose locations. 18 D Real Data: Additional Visualizations Figure 5: Comparing true vs. reconstructed ...

  40. [1999]

    URL http://www.jstor.org/stable/2680726

    ISSN 13697412, 14679868. URL http://www.jstor.org/stable/2680726

  41. [2005]

    URL http://jmlr.org/papers/v6/quinonero-candela05a.html

  42. [2017]

    URL https://openreview.net/forum?id=Sy2fzU9gl

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.