Pith. sign in

REVIEW 3 major objections 4 minor 30 references

Enhancing Neural Autoregressive Distribution Estimators for Image Reconstruction

T0 review · 3 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read Observing an evenly spread (low-discrepancy) subset of pixels, instead of a random patch, yields lower test loss, faster convergence, and sharper reconstructions from a lightweight autoregressive image model.

desk verdict A useful lightweight model and a neat Sobol-patch idea, but the central LD-vs-random comparison is not yet controlled enough to support the headline claim. read the letter →

arxiv 2506.05391 v2 pith:I7MXUFSK submitted 2025-06-03 eess.IV cs.CVcs.LGstat.AP

Ambrose Emmett-Iwaniw , Nathan Kirk This is my paper
classification eess.IVcs.CVcs.LGstat.AP
keywords autoregressivemodelsimagereconstructionlow-discrepancysamplingConvNADEBetadistributionSobol'sequencepixelpatchesquasi-MonteCarlo
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

Autoregressive density estimators factor an image's likelihood into a chain of pixel-by-pixel conditionals, so the subset of pixels the model conditions on shapes what it learns. This paper studies a practical variant: reconstruct the full image from a fixed small subset of observed pixels, a 'pixel patch'. The authors propose ConvNADE-Beta-Color, a convolutional NADE adapted to real-valued color images by modeling each color channel's conditional distribution as a Beta distribution, and train it to predict the unobserved pixels from the patch. Their central finding across four benchmark datasets is that a low-discrepancy patch—pixels chosen deterministically from a Sobol' sequence so they spread evenly over the image—produces lower test loss, faster convergence, and clearer reconstructions than a uniformly random patch of the same size. If correct, the result says that whenever an acquisition or compression system controls which pixels are kept, spreading them evenly improves reconstruction quality at no extra cost.

What carries the argument

The load-bearing mechanism is the low-discrepancy pixel patch: a deterministic subset of $2^k$ pixels obtained by mapping the first $2^{2m}$ points of a two-dimensional Sobol' sequence onto a $2^m \times 2^m$ image grid, with the pixel index for each point given by $\lfloor 2^m u_n^{(1)} \rfloor + 2^m (2^m - 1 - \lfloor 2^m u_n^{(2)} \rfloor) + 1$. Because the Sobol' sequence is a digital net, its early points are stratified evenly over the unit square, so the resulting patch covers the image more uniformly than independent uniform random draws. The paper pairs this patch with ConvNADE-Beta-Color, a convolutional neural autoregressive model whose final layer outputs, for each of the three color channels, the shape and scale parameters of a Beta distribution; the mask indicating which pixels are observed is concatenated as an extra input channel to disambiguate true zeros from masked zeros. Training minimizes the negative log-likelihood of the unobserved pixels conditioned on the patch.

What would settle it

Train the model on a single random patch that is sampled once and held fixed for the whole run, so the only difference from the evenly spread condition is uniformity; if the gap disappears or flips, the benefit comes from a fixed rather than resampled patch, not from even coverage.

Watch

Extended reading notes

Core claim

The paper's central discovery is that the spatial distribution of the observed pixel patch is a first-order factor in how well a masked conditional autoregressive model reconstructs an image. Training ConvNADE-Beta-Color on a low-discrepancy patch—a set of $2^k$ pixels whose coordinates are taken from the first $2^{2m}$ points of a two-dimensional Sobol' sequence mapped onto a $2^m \times 2^m$ grid—consistently outperforms training on a uniformly random patch of the same size. Table 1 reports lower test loss for the low-discrepancy patch on all four datasets (Binarized MNIST, FER2013, CIFAR-10, LHQ), and Figure 3 shows training and validation curves that converge faster and to lower values. Reconstructions in Appendix A are sharper and show more detailed object features under the low-discrepancy patch. The authors interpret this as evidence that, where the observation pattern can be designed (for example in single-pixel cameras or superpixel-based compression), pixels should be sampled or stored to preserve uniform coverage.

Load-bearing premise

The load-bearing premise is that the only difference between the random and evenly spread patches is spatial uniformity, so if the random patch is resampled during training while the evenly spread patch is fixed, the reported advantage could come from a stable training target rather than from even coverage.

Editorial extensions

If this is right

  • In any acquisition or compression setting where the observed pixel set can be designed—single-pixel cameras, compressive sampling, superpixel-based storage—a low-discrepancy pattern should beat random sampling at the same pixel budget.
  • ConvNADE-Beta-Color offers a lightweight, single-GPU-trainable alternative to high-capacity inpainting models, and its single Beta output trains faster than a 10-component Gaussian mixture (2375 versus 3221 seconds on FER2013) while removing the mixture-count hyperparameter.
  • Because the same fixed-patch training procedure improves the original binary ConvNADE on Binarized MNIST as well, the uniformity benefit is not specific to the Beta output layer.
  • The effect appears across four datasets with different content (digits, faces, objects, landscapes), so the advantage of uniform coverage is likely general rather than tied to one data type.

Reading between the lines

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

  • Because the random-patch condition averages over five random patches while the low-discrepancy patch is a single fixed set, part of the measured gap could come from conditioning on a stationary pattern rather than from uniformity itself; holding one random patch fixed throughout training would isolate the uniformity effect.
  • A direct application would be lossy compression: store the pixels of a Sobol' patch and their coordinates, reconstruct the rest with the trained model; the results suggest this will outperform storing the same number of random pixels at equal bit cost.
  • For programmable sensors, a fixed quasirandom pixel schedule is a memoryless design that avoids per-image optimization, so the principle transfers to hardware with toggleable pixels such as event cameras or programmable apertures.
  • Images with highly localized information, such as dense text or faces, might favor content-adaptive patches over uniform ones; the paper's benchmarks do not test that regime, so the uniform recommendation is not obviously optimal there.
Share X Bluesky LinkedIn Reddit HN

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 introduces ConvNADE-Beta-Color, an extension of the convolutional neural autoregressive distribution estimator that models each RGB channel with a Beta distribution and is trained with a masked conditional objective for image reconstruction from a fixed subset of P observed pixels. The paper then compares random pixel patches with low-discrepancy (Sobol-based) pixel patches, reporting that low-discrepancy patches yield lower test loss, faster convergence, and visually clearer reconstructions on Binarized MNIST, FER2013, CIFAR-10, and LHQ. The central experimental claim is that uniform-coverage observation patterns improve reconstruction quality.

Significance. If the comparison between random and low-discrepancy patches were controlled, the finding would be practically relevant for applications where the observation pattern is designable, such as single-pixel cameras and superpixel compression. The proposed model is a lightweight alternative to high-capacity generative models and the authors state that code and data are available. The paper also includes a concrete training-time comparison between the Beta-output model and a Gaussian-mixture-output variant. However, the current evidence for the headline claim is confounded by experimental-design issues, and the description of the low-discrepancy patch construction contains an internal inconsistency that prevents reproduction of the exact patches used in the experiments.

major comments (3)
  1. [3.2 and 4.1] The definition of low-discrepancy pixel patches is inconsistent with the image sizes and patch sizes used in the experiments. Section 3.2 states that "all images used in our experiments in Section 4 are rescaled to 25×25" and defines the patch for an image of dimension 2^m × 2^m with patch size 2^k for k ≤ m. Section 4.1 instead states "All images are resized to 32×32," and Figure 2 also shows a 32×32 image. With a 32×32 image, m = 5, so the condition k ≤ m would restrict the patch size to at most 32 pixels, yet the experiments use pixel patches of size 128. The construction as written therefore does not generate the reported experimental patches, and the paper must clarify the intended image size, the correct constraint on k, and how the 128-pixel Sobol patch is obtained.
  2. [4.3, Table 1, and Figure 3] The random-versus-low-discrepancy comparison is not sufficiently controlled. The paper reports the mean test loss over five random patches with 95% confidence intervals, but gives a single deterministic Sobol patch with no variance or error bars. It is never stated whether each random-patch run uses one fixed mask for all training epochs or resamples the mask every minibatch. If resampling occurs, the random-patch task is non-stationary while the low-discrepancy task is stationary, so the observed gap could be due to optimization ease rather than pixel coverage. Even if each random-patch run uses a fixed mask, the comparison relies on one particular Sobol set, so the result may reflect the specific pixel positions rather than the low-discrepancy property. The authors should specify the exact protocol and report low-discrepancy variance, for example by using multiple scrambled nets or randomized starts of the Sobol sequence.
  3. [3.2 (objective function)] The loss in Equation for L_{b,P,Beta} is averaged only over the complement of the observed patch S_P^•. Because the random and low-discrepancy conditions use different observed sets, they are evaluated on different target pixel sets. A lower negative log-likelihood under the low-discrepancy condition could therefore reflect that the particular unobserved pixels in the low-discrepancy case are easier to predict, rather than that the observed low-discrepancy patch is more informative. To isolate the effect of the observed set, the evaluation should hold the target pixel set fixed across conditions, or average over all possible target sets in a symmetric way.
minor comments (4)
  1. [3.1] The output distribution parameters are described as "shape and scale" for the Beta distribution, but the Beta distribution is usually parameterized by two shape parameters (α and β); "scale" is nonstandard and should be renamed to avoid confusion. The paper should also state how pixel values that are exactly 0 or 1 are handled, since the Beta density is zero at the boundaries and the training data are real-valued images in [0,1].
  2. [3.2] The mask M_{o<d} is inherited from the autoregressive formulation, but in the fixed-subset reconstruction setting there is no ordering. The paper should explicitly define the mask as the indicator of the observed set S_P^• and explain how this mask is combined with the input image across the color channels.
  3. [4.1 and Figure 3 caption] The text says "the mean loss over five random orderings is given" for the random patch, while Table 1 states "mean average loss over 5 random patches with 95% CI." These phrasings are inconsistent; the authors should clarify whether the five runs differ in the random patch mask, the ordering, or both.
  4. [4.3] The model contribution is validated mainly through reconstruction-quality comparisons, but there is no baseline comparison for density estimation quality against other real-valued autoregressive models such as RNADE or ConvNADE-MoG on the same datasets. Adding such a comparison, even on one dataset, would strengthen the claim that the Beta-output model is a generally useful variant.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: the LD-versus-random finding is an empirical comparison, not an equation-level reduction to the paper's own inputs.

full rationale

The paper makes no theoretical prediction that is fitted to data and then renamed as a result. Its central claim—that low-discrepancy pixel patches yield lower test loss, faster convergence, and clearer reconstructions than random patches—is an experimental outcome measured on held-out pixels (Section 4.3, Table 1, Figure 3). The low-discrepancy patch is constructed deterministically from the Sobol' sequence, whose discrepancy properties are imported from external quasi-Monte Carlo literature (Sobol' [22], Dick and Pillichshammer [7]); the paper does not derive those properties from its own results. The loss function in Section 3.2 is a standard negative log-likelihood over unobserved pixels, and no parameter of the LD-vs-random comparison is fitted to the test-set numbers that support the claim. There are no load-bearing self-citations: the references to NADE, ConvNADE, RNADE, and QMC software are to independent prior work, not to the present authors' own unpublished results. The only passages resembling limitations are the acknowledgment that the uni-modal Beta distribution may limit expressiveness (Section 3) and the experimental-design caveat that the random-patch condition is averaged over five patches while the LD condition is a single fixed Sobol' set, with the loss evaluated on different complement pixel sets. Those are potential threats to experimental control and generality, but they are not circularity: the LD advantage does not reduce by construction to the model's inputs, and no equation in the paper defines the LD result in terms of the loss or vice versa. Accordingly, the derivation chain is self-contained as an empirical study, and no circular step can be exhibited with a specific reduction.

Assumptions & free parameters 5 free parameters · 3 assumptions · 0 invented entities

The central claim is empirical; no new entities or fitted constants are introduced. The model uses standard probabilistic and QMC tools. The listed free parameters are experimental hyperparameters chosen by hand, not constants fitted to make a derivation work.

free parameters (5)
  • Pixel patch size P = 128
    Chosen by hand for all experiments; the central comparison between random and LD patches uses this fixed value.
  • Learning rate = 1e-4
    Described as 'found to be the optimal choice for training'; no search details are given.
  • Number of convolutional layers L = 8 (from Figure 1 caption)
    Architecture choice; not justified by experiments or ablation.
  • Dropout probability = 0.5
    Set between hidden layers to avoid overfitting; no sensitivity analysis.
  • Batch size = 100
    Fixed for all experiments without justification.
assumptions (3)
  • standard math Chain rule of probability factorizes the joint density into a product of conditionals.
    Used in Equation (1), Section 1.2, as the foundation of autoregressive density estimation.
  • standard math The Sobol' sequence is a low-discrepancy digital net with stratification properties.
    Invoked in Section 3.2 to argue that the resulting pixel patch has uniform coverage; property taken from QMC literature [7,22].
  • domain assumption A Beta distribution adequately models real-valued pixel intensities in [0,1].
    Adopted in Section 3.1; boundary behavior at 0 and 1 is not discussed, and no comparison to mixture alternatives is made for reconstruction quality.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Enhancing Neural Autoregressive Distribution Estimators for Image Reconstruction." pith.science (2026). https://pith.science/paper/I7MXUFSK

@misc{pith2026250605391,
  author       = {Pith},
  title        = {Pith review of: Enhancing Neural Autoregressive Distribution Estimators for Image Reconstruction},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/I7MXUFSK}},
  note         = {Machine review of arXiv:2506.05391}
}
abstract

Autoregressive models are often employed to learn distributions of image data by decomposing the $D$-dimensional density function into a product of one-dimensional conditional distributions. Each conditional depends on preceding variables (pixels, in the case of image data), making the order in which variables are processed fundamental to the model performance. In this paper, we study the problem of observing a small subset of image pixels (referred to as a pixel patch) to predict the unobserved parts of the image. As our prediction mechanism, we propose a generalized version of the convolutional neural autoregressive distribution estimation (ConvNADE) model adapted for real-valued and color images. Moreover, we investigate the quality of image reconstruction when observing both random pixel patches and low-discrepancy pixel patches inspired by quasi-Monte Carlo theory. Experiments on benchmark datasets demonstrate that, where design permits, pixels sampled or stored to preserve uniform coverage improves reconstruction fidelity and test performance.

Figures

Figures reproduced from arXiv: 2506.05391 by the authors.

Figure 1
Figure 1. ConvNADE and ConvNADE-Beta-Color with 𝐿 = 8 convolutional layers, each defined as channels @ filter size (conv type). For ConvNADE, the image has 2 channels: one for the binary pixel value and one for the mask, and outputs a 1 layer Bernoulli parameter map. For ConvNADE-Beta-Color, the image has 4 channels: one for each color and one for the mask and the output channel has 6 layers representing Beta shape and scale … view at source ↗
Figure 2
Figure 2. A 32×32 image from LHQ illustrating pixel patches. (Left) A random patch and (Right) a low-discrepancy patch of size 128. 3.2 Training ConvNADE-Beta-Color for Image Reconstruction As previously noted, ConvNADE models are traditionally trained on a random subset of pixel orderings sampled from the full set of permutations, D!. This corresponds to conditioning on one pixel at a time in a fully autoregressive manner. H… view at source ↗
Figure 3
Figure 3. For each data set, the left and right panels show the training and validation loss [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Reconstructed images from binarized MNIST using the ConvNADE model. ( [PITH_FULL_IMAGE:figures/full_fig_p014_4.png]
Figure 5
Figure 5. Figure 5: Reconstructed images from FER2013 using the ConvNADE-Beta-Color model [PITH_FULL_IMAGE:figures/full_fig_p014_5.png]
Figure 6
Figure 6. Figure 6: Reconstructed images from CIFAR-10 using the ConvNADE-Beta-Color model [PITH_FULL_IMAGE:figures/full_fig_p015_6.png]
Figure 7
Figure 7. Figure 7: Reconstructed images of LHQ using the ConvNADE-Beta-Color model ( [PITH_FULL_IMAGE:figures/full_fig_p015_7.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

30 extracted references · 24 canonical work pages

  1. [1]

    Signal Processing 212, 109158 (2023)

    Asante-Mensah, M.G., Phan, A.H., Ahmadi-Asl, S., Aghbari, Z.A., Cichocki, A.: Image reconstruction using superpixel clustering and tensor completion. Signal Processing 212, 109158 (2023). DOI https://doi.org/10.1016/j.sigpro.2023.109158

    work page arXiv 2023

  2. [2]

    Advances in Neural Information Processing Systems 12 (1999)

    Bengio, Y., Bengio, S.: Modeling high-dimensional discrete data with multi-layer neural networks. Advances in Neural Information Processing Systems 12 (1999)

    work page 1999

  3. [3]

    IEEE Transactions on Image Processing 19(10), 2634–2645 (2010)

    Bugeau, A., Bertalm ´ıo, M., Caselles, V., Sapiro, G.: A comprehensive framework for image inpainting. IEEE Transactions on Image Processing 19(10), 2634–2645 (2010)

    work page 2010

  4. [4]

    In: International conference on machine learning, pp

    Chen, M., Radford, A., Child, R., Wu, J., Jun, H., Luan, D., Sutskever, I.: Generative pretrain- ing from pixels. In: International conference on machine learning, pp. 1691–1703. PMLR (2020)

    work page 2020

  5. [5]

    Quasi-Monte Carlo Software

    Choi, S.C.T., Hickernell, F.J., Jagadeeswaran, R., McCourt, M.J., Sorokin, A.G.: Quasi-Monte Carlo software (2021). URL https://arxiv.org/abs/2102.07833

    work page Pith review arXiv 2021

  6. [6]

    In: Proceedings of the 35th International Conference on Neural Information Processing Systems, NIPS ’21

    Dhariwal, P., Nichol, A.: Diffusion models beat gans on image synthesis. In: Proceedings of the 35th International Conference on Neural Information Processing Systems, NIPS ’21. Curran Associates Inc., Red Hook, NY, USA (2021)

    work page 2021

  7. [7]

    Cambridge University Press (2010)

    Dick, J., Pillichshammer, F.: Digital nets and sequences: discrepancy theory and quasi–Monte Carlo integration. Cambridge University Press (2010)

    work page 2010

  8. [8]

    In: Proceedings of the AAAI Conference on Artificial Intelligence, vol

    Dong, X., Bao, J., Zhang, T., Chen, D., Zhang, W., Yuan, L., Chen, D., Wen, F., Yu, N., Guo, B.: Peco: Perceptual codebook for bert pre-training of vision transformers. In: Proceedings of the AAAI Conference on Artificial Intelligence, vol. 37, pp. 552–560 (2023)

    work page 2023

Show all 30 references
  1. [9]

    IEEE Signal Processing Magazine 25(2), 83–91 (2008)

    Duarte, M.F., Davenport, M.A., Takhar, D., Laska, J.N., Sun, T., Kelly, K.F., Baraniuk, R.G.: Single-pixel imaging via compressive sampling. IEEE Signal Processing Magazine 25(2), 83–91 (2008). DOI 10.1109/MSP.2007.914730

    2008

  2. [10]

    MIT press (1998)

    Frey, B.J.: Graphical models for machine learning and digital communication. MIT press (1998)

    1998

  3. [11]

    In: Neural information processing: 20th international conference, ICONIP 2013, daegu, korea, november 3-7, 2013

    Goodfellow, I.J., Erhan, D., Carrier, P.L., Courville, A., Mirza, M., Hamner, B., Cukierski, W., Tang, Y., Thaler, D., Lee, D.H., et al.: Challenges in representation learning: A report on three machine learning contests. In: Neural information processing: 20th international c...

    2013

  4. [12]

    In: 2006 International Conference on Image Processing, pp

    Han, S., Vasconcelos, N.: Image compression using object-based regions of interest. In: 2006 International Conference on Image Processing, pp. 3097–3100 (2006). DOI 10.1109/ICIP. 2006.313095

    2006

  5. [13]

    In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp

    He, K., Chen, X., Xie, S., Li, Y., Doll ´ar, P., Girshick, R.: Masked autoencoders are scalable vision learners. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp. 16000–16009 (2022) NADE for Image Reconstruction 13

    2022

  6. [14]

    ACM computing surveys (CSUR) 54(10s), 1–41 (2022)

    Khan, S., Naseer, M., Hayat, M., Zamir, S.W., Khan, F.S., Shah, M.: Transformers in vision: A survey. ACM computing surveys (CSUR) 54(10s), 1–41 (2022)

    2022

  7. [15]

    arXiv preprint arXiv:1412.6980 (2014)

    Kingma, D.P.: Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980 (2014)

    2014 arXiv

  8. [16]

    Krizhevsky, A., Hinton, G., et al.: Learning multiple layers of features from tiny images (2009)

    2009

  9. [17]

    In: Proceedings of the fourteenth international conference on artificial intelligence and statistics, pp

    Larochelle, H., Murray, I.: The neural autoregressive distribution estimator. In: Proceedings of the fourteenth international conference on artificial intelligence and statistics, pp. 29–37. JMLR Workshop and Conference Proceedings (2011)

    2011

  10. [18]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp

    Lugmayr, A., Danelljan, M., Romero, A., Yu, F., Timofte, R., Van Gool, L.: Repaint: In- painting using denoising diffusion probabilistic models. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp. 11461–11471 (2022)

    2022

  11. [19]

    In: 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp

    Rombach, R., Blattmann, A., Lorenz, D., Esser, P., Ommer, B.: High-Resolution Image Synthesis with Latent Diffusion Models . In: 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp. 10674–10685. IEEE Computer Society, Los Alamitos, CA, USA (2022). DO...

    2022

  12. [20]

    In: Proceedings of the 25th international conference on Machine learning, pp

    Salakhutdinov, R., Murray, I.: On the quantitative analysis of deep belief networks. In: Proceedings of the 25th international conference on Machine learning, pp. 872–879 (2008)

    2008

  13. [21]

    In: Proceedings of the IEEE/CVF international conference on computer vision, pp

    Skorokhodov, I., Sotnikov, G., Elhoseiny, M.: Aligning latent and image spaces to connect the unconnectable. In: Proceedings of the IEEE/CVF international conference on computer vision, pp. 14144–14153 (2021)

    2021

  14. [22]

    USSR Computational Mathematics and Mathematical Physics 7(4), 86–112 (1967)

    Sobol’, I.: On the distribution of points in a cube and the approximate evaluation of integrals. USSR Computational Mathematics and Mathematical Physics 7(4), 86–112 (1967)

    1967

  15. [23]

    The journal of machine learning research 15(1), 1929–1958 (2014)

    Srivastava, N., Hinton, G., Krizhevsky, A., Sutskever, I., Salakhutdinov, R.: Dropout: a simple way to prevent neural networks from overfitting. The journal of machine learning research 15(1), 1929–1958 (2014)

    2014

  16. [24]

    Journal of Machine Learning Research 17(205), 1–37 (2016)

    Uria, B., C ˆot´e, M.A., Gregor, K., Murray, I., Larochelle, H.: Neural autoregressive distribution estimation. Journal of Machine Learning Research 17(205), 1–37 (2016)

    2016

  17. [25]

    Advances in Neural Information Processing Systems 26 (2013)

    Uria, B., Murray, I., Larochelle, H.: Rnade: The real-valued neural autoregressive density- estimator. Advances in Neural Information Processing Systems 26 (2013)

    2013

  18. [26]

    In: International Conference on Machine Learning, pp

    Uria, B., Murray, I., Larochelle, H.: A deep and tractable density estimator. In: International Conference on Machine Learning, pp. 467–475. PMLR (2014)

    2014

  19. [27]

    In: Proceedings of the 25th international conference on Machine learning, pp

    Vincent, P., Larochelle, H., Bengio, Y., Manzagol, P.A.: Extracting and composing robust features with denoising autoencoders. In: Proceedings of the 25th international conference on Machine learning, pp. 1096–1103 (2008)

    2008

  20. [28]

    Journal of machine learning research 11(12) (2010)

    Vincent, P., Larochelle, H., Lajoie, I., Bengio, Y., Manzagol, P.A., Bottou, L.: Stacked denois- ing autoencoders: Learning useful representations in a deep network with a local denoising criterion. Journal of machine learning research 11(12) (2010)

    2010

  21. [29]

    In: Proceedings of the 32nd ACM International Conference on Multimedia, MM ’24, p

    Wu, H., He, L., Zhang, M., Chen, D., Luo, K., Luo, M., Zhou, J.Z., Chen, H., Lv, J.: Diffusion posterior proximal sampling for image restoration. In: Proceedings of the 32nd ACM International Conference on Multimedia, MM ’24, p. 214–223. Association for Computing Machinery, Ne...

    2024

  22. [30]

    Yule, G.U.: On a method of investigating periodicities disturbed series, with special reference to wolfer’s sunspot numbers. Phil. Trans. A pp. 267–298 (1927) Notes and Comments. The first author is supported by NSERC Discovery Grant RGPIN- 238959. The second author is support...

    1927

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.