Pith. sign in

REVIEW 3 major objections 5 minor 42 references

PolyGAN: High-Order Polynomial Generators

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

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

Pith's one-line read The paper establishes that a GAN generator built as a high-order polynomial with coupled tensor decompositions can approximate data distributions without activation functions and improve over standard GAN generators.

desk verdict Solid tensor-decomposition core and a clean network-equivalence result; the 'no activations' headline only holds for low-dimensional synthetic data and the SOTA gains are from nonlinear PolyGAN variants, so the abstract oversells but the paper deserves referee attention. read the letter →

arxiv 1908.06571 v2 pith:XEQJT7FN submitted 2019-08-19 cs.LG stat.ML

Grigorios Chrysos , Stylianos Moschoglou , Yannis Panagakis , Stefanos Zafeiriou This is my paper · ORCID
classification cs.LGstat.ML
keywords generativeadversarialnetworkshigh-orderpolynomialstensordecompositioncoupledcanonicalpolyadicgeneratorwithoutactivationfunctionsInceptionScoreFrechetDistanceimagegeneration
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 tries to establish that the generator in a GAN can be a high-order polynomial of the latent vector instead of a hierarchy of activated linear maps, and that this change helps rather than hurts generation quality. Because an $N$th-order polynomial has exponentially many coefficients, the authors write the coefficient tensors through two coupled canonical polyadic decompositions that share factors across orders, and they prove the resulting factorized models collapse into the hierarchical linear/convolutional architectures used in practice. On this basis they report that a PolyGAN generator with no activation functions can synthesize digits and faces, and that PolyGAN improves Inception Score and FID over DCGAN, SNGAN, and SAGAN. If the claim holds, GAN generators become interpretable as truncated Taylor expansions and polynomial approximation theory becomes applicable to the study of adversarial training.

What carries the argument

The load-bearing object is the coupled CP decomposition: each $n$th-order coefficient tensor of the polynomial is written as a sum of rank-one tensors, with the factor matrices belonging to lower-order terms reused by higher-order terms (e.g., $W^{[1]}=CU_{[1]}^T$, $W^{[2]}_{(1)}=C(U_{[3]}\odot U_{[1]})^T+C(U_{[2]}\odot U_{[1]})^T$, and so on). The identity that lets these factorized tensors execute as networks is the Khatri-Rao/Hadamard rule $\left(\bigodot_{\nu=1}^N A_\nu\right)^T\left(\bigodot_{\nu=1}^N B_\nu\right)= (A_1^T B_1)\ast\cdots\ast(A_N^T B_N)$, which converts a product of many factors into a sequence of Hadamard products with small linear maps. Algorithms 1 and 2 turn this identity into hierarchical generators: the nested variant computes $\kappa=((B^{[1]})^T b^{[1]})\ast((A^{[1]})^T v)$ and then iterates $\kappa=(S^{[n]}\kappa+(B^{[n]})^T b^{[n]})\ast((A^{[n]})^T v)$, finishing with $x=\beta+C\kappa$, so each added polynomial order is one extra Hadamard-plus-linear block.

What would settle it

Train a PolyGAN with small rank $k$ on a synthetic distribution that is, by construction, a high-rank polynomial generator (e.g., $G^*(z)=\sum_{r=1}^{1000} c_r (a_r^\top z)^4$ with random $a_r$); if full-rank polynomial generators fit it but PolyGAN's FID stays far above the full model even as the order and width grow, the coupled low-rank assumption fails and the image results are dataset-specific rather than a universal property.

Watch

Extended reading notes

Core claim

The central claim is that universal approximation by GAN generators does not require elementwise nonlinearities: a generator $G(z)=\beta+\sum_{n=1}^N \mathcal{W}^{[n]}\prod_{j=2}^{n+1}\times_j z$ with tensor-valued weights can serve as a data generator, provided the tensors $\mathcal{W}^{[n]}$ are coupled through a CP-type decomposition with shared factors. The paper introduces two such decompositions, the coupled CP decomposition and the coupled nested CP decomposition, and proves (Claims 1–4) that each is exactly reproduced by a stack of linear or convolutional blocks with Hadamard products, so the polynomial can be trained end-to-end as an ordinary GAN generator. The authors state that this gives the first demonstration of a GAN generator approximating a data distribution without activation functions, and support it by experiments on analytic 2D/3D manifolds, MNIST digits, YaleB faces, CIFAR-10, and ImageNet, where PolyGAN improves Inception Score and FID over the corresponding DCGAN, SNGAN, and SAGAN baselines.

Load-bearing premise

The entire construction rests on the assumption that the polynomial coefficient tensors for real image generators lie close to the low-rank, shared-factor family of the coupled CP decomposition; the proofs only show the architectures compute the polynomial, not that this restricted class can approximate every distribution a standard GAN can fit.

Editorial extensions

If this is right

  • A generator built solely from linear/convolutional blocks can learn nontrivial image distributions: the paper reports plausible digits and faces with a single output tanh and no hidden activations.
  • Adding one polynomial order adds one hierarchical block, so PolyGAN offers an architecture-independent knob for model capacity: the paper shows that removing the fourth-order term removes fine detail.
  • PolyGAN improves Inception Score and FID over DCGAN, SNGAN, and SAGAN with roughly a 2–3% increase in generator parameters, whereas concatenating noise at each block adds many more parameters without the same gain.
  • The polynomial structure carries over to arbitrary decoder designs: replacing the standard generator in DCGAN, SNGAN, or SAGAN by its polynomial counterpart improves the reported scores.
  • Because no activations are required, the entire generator is a polynomial map, so existing machinery for polynomial approximation, not just neural-network training, can be applied to GANs.

Reading between the lines

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

  • The coupled-CP parameterization makes the generator linear in its final-layer coefficients $C$, which suggests that convergence analyses for deep linear networks (in the regime where the factors are fixed or convexified) might transfer to GAN training; the paper does not make this claim.
  • The expressivity that matters is that of the low-rank shared-factor family, not of arbitrary polynomials. A high-rank interaction among the latent dimensions would be exactly representable by an unconstrained polynomial but only approximately by PolyGAN at fixed rank $k$, so performance on a given dataset depends on how much of the true generator's mass sits in the coupled low-rank subspace.
  • A testable consequence of interpreting StyleGAN's gains as polynomial refinements rather than style transfer would be that the same base generator should match StyleGAN's image-quality curve as the polynomial order increases, without adaptive instance normalization; the authors suggest the interpretation but do not run this comparison.
  • Since every generated pixel is an analytic function of the latent vector, PolyGAN cannot represent distributions supported on lower-dimensional manifolds with sharp boundaries exactly; near such boundaries accuracy should improve only as polynomial order grows, which is a falsifiable prediction the paper does not address.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper proposes PolyGAN, a GAN generator expressed as an N-th order multivariate polynomial in the latent code z, with the polynomial parameters organized as tensors. Two coupled CP-type decompositions with shared factors are introduced, and the authors prove (Lemmas 1-2 and Claims 1-4) that the factorized expansions can be implemented as hierarchical networks built from linear/convolutional blocks and Hadamard products (Algorithms 1-2). Experiments cover a 2D sinusoid, several 3D manifolds, MNIST/CIFAR/YaleB generation with linear blocks plus a final tanh, and CIFAR/ImageNet generation using DCGAN/SNGAN/SAGAN backbones modified with multiplicative noise injections; the latter are reported as outperforming the baselines in Inception Score and FID. The paper also compares the two proposed decompositions qualitatively in an appendix.

Significance. If the claims are appropriately scoped, the paper makes a useful conceptual contribution by connecting polynomial expansions, coupled tensor decompositions, and hierarchical linear networks. The algebraic equivalence proofs are explicit and checkable, and the experimental section reports mean and standard deviation over 10 runs and includes parameter counts, which is a strength. However, the headline claims currently outrun the evidence: the no-activation result is only demonstrated on low-dimensional manifolds, and the state-of-the-art comparisons use nonlinear generators that are not polynomial maps. The core derivation appears sound, but the paper needs a careful reframing before publication.

major comments (3)
  1. [Abstract; Section 4.2; Appendix E] The abstract claims that 'a GAN generator can approximate the data distribution without using any activation functions,' but the image experiments use a tanh at the generator output. Section 4.2 states 'We only add a tanh in the output of the generator for normalization purposes,' and Figure 5 is captioned 'with a single activation in the generator.' A tanh is an activation function, and a tanh composed with a polynomial is not a polynomial map. The no-activation claim is therefore established only for the low-dimensional synthetic setups of Sections 4.1, D, and G; the abstract and contribution 3 should be revised to state that the no-activation result applies to those settings, not to high-dimensional image generation.
  2. [Section F; Algorithms 3-4; Tables 11-12] The state-of-the-art comparisons do not test polynomial generators. In Algorithm 4, v = phi(Linear(z)) with phi = ReLU, h = phi(Linear(v)), and the original nonlinear residual blocks are retained. This is a multiplicative-feature-modulation variant of SNGAN, not the polynomial generator of Eqs. (7) and (10), and it is not equivalent to any polynomial map. Consequently, the IS/FID improvements over DCGAN, SNGAN, and SAGAN in Tables 11 and 12 cannot be attributed to the polynomial expansion; they may be due to the noise-injection mechanism alone. The authors should either rename these variants (e.g., 'modulated SNGAN') or provide an ablation that isolates the polynomial structure from the multiplicative injection mechanism before claiming that PolyGAN improves upon these architectures.
  3. [Section 2.2; Eqs. (5)-(10)] The Stone-Weierstrass motivation applies to the full polynomial model in Eq. (5), while the implemented models are the low-rank coupled CP subclasses of Eqs. (6), (7), and (10). No approximation-error or expressivity analysis is provided for the rank-k factors shared across orders; the statement 'assume that the weights exhibit redundancy and hence the parameter tensors are of low-rank' is an assumption, not a guarantee. The abstract's 'universal function approximator' language should therefore be limited to the full polynomial or explicitly qualified as an empirical hypothesis for the low-rank coupled form.
minor comments (5)
  1. [Section 4.2; Appendix E; Figure 5] The section title 'Digit Generation with Linear Blocks' is contradicted by Figure 5's caption 'with a single activation in the generator' and by Appendix E's statement that there is 'a single activation function only on the output (i.e., tanh)'. Please make the wording consistent.
  2. [Section F.4] The SAGAN comparison is run for 400,000 iterations with batch size 64 instead of the original batch size 256, so the baseline is under-trained relative to the published SAGAN results; the abstract's 'state of the art' should be qualified with these computational budgets.
  3. [Tables 11-12] Standard deviations are reported but no significance tests are given; several FID intervals overlap (e.g., conditional SNGAN FID 14.70 +/- 0.97 vs. 13.52 +/- 0.60 in Table 12), so the improvement claim should be tempered or accompanied by paired comparisons across the 10 seeds.
  4. [Section F.1; References] There is a typo in Section F.1 ('WThe rest details remain the same'), the word 'acrhitectures' appears in Section F, and the reference list duplicates the Sidiropoulos et al. (2017) entry; these should be corrected.
  5. [Appendix G] The comparison between the Coupled CP and Coupled nested CP decompositions is qualitative; reporting quantitative metrics (e.g., IS/FID or a distributional distance) would support the claim that 'they perform similarly.'

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the generator-to-architecture equivalence is a direct algebraic derivation, and the empirical gains are benchmark comparisons rather than fitted quantities relabeled as predictions.

full rationale

The paper's central derivation (Eq. 5 to Eqs. 7/10 and Algorithms 1/2) is an algebraic identity: applying Lemma 2, the Khatri-Rao/Hadamard identity, to the CP-factorized polynomial expansion yields the hierarchical linear-block implementation. The coupled low-rank factorizations are introduced explicitly as modeling assumptions, not as consequences of the experiments, and the appendix proves the network recursion from the polynomial form by induction (Claim 4). No fitted parameter is renamed as a prediction: the IS/FID tables compare trained PolyGAN variants against re-run baselines, and choices such as polynomial order and rank are engineering settings reported in the experimental section, not outputs of the derivation. The only load-bearing external premise, Stone-Weierstrass, is used to motivate polynomials generally, not to prove the specific result. One concern belongs to correctness rather than circularity: the abstract's claim of 'without using any activation functions' is not supported by the image experiments, since Section 4.2 and Appendix E/G add a final tanh (an activation), and Section F uses ReLU and nonlinear residual blocks; a tanh composed with a polynomial is not a polynomial map. That inconsistency does not make the derivation circular.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central claim rests on a small set of hyperparameters (polynomial order N, tensor rank or width, and learned scaling vectors) and on several unproved premises: transfer of Stone-Weierstrass to the restricted low-rank architecture, existence of low-rank structure in the generator weights, and stable GAN training. No new physical entities are introduced. The mathematical derivation linking polynomial expansion to network architecture is the main contributed content.

free parameters (3)
  • Polynomial order N = N=12 (sinusoid), 8/12 (model comparison), 4 (MNIST), 5 (YaleB/CIFAR10)
    Chosen per experiment by hand; the approximation power and number of terms scale with N, so the central claim depends on selecting it.
  • Decomposition rank / layer width k = Width 15 for sinusoid; 256 channels for MNIST/CIFAR; residual block widths for ImageNet
    The coupled CP rank is set by network width. No criterion or theorem fixes a sufficient rank for the target distribution.
  • Learnable scaling vectors b[n] = Learned, values not reported
    Introduced in Equation (8) as scaling factors; they are extra fitted parameters not present in the raw polynomial expansion (5).
assumptions (4)
  • standard math Any continuous function on a closed interval can be uniformly approximated by polynomials (Stone-Weierstrass).
    Invoked in the Introduction and Section 2.2 to motivate polynomial generators; it does not by itself cover the restricted low-rank coupled polynomial subclass used in PolyGAN.
  • domain assumption The parameter tensors W[n] admit low-rank coupled CP decompositions with shared factors.
    Section 2.2 assumes the weights exhibit redundancy and are low-rank; this is central to the parameter reduction and to the specific network architectures.
  • domain assumption GAN adversarial training reaches a useful equilibrium for the proposed generator.
    All empirical results depend on GAN training dynamics; the paper does not analyze convergence or stability for PolyGAN specifically.
  • domain assumption Inception Score and FID are valid proxies for generation quality.
    Section F uses IS and FID for quantitative comparison, acknowledging their known shortcomings; the central empirical claims rely on these metrics.

how reviews work

0 comments
Cite this review

Pith. "Pith review of PolyGAN: High-Order Polynomial Generators." pith.science (2026). https://pith.science/paper/XEQJT7FN

@misc{pith2026190806571,
  author       = {Pith},
  title        = {Pith review of: PolyGAN: High-Order Polynomial Generators},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/XEQJT7FN}},
  note         = {Machine review of arXiv:1908.06571}
}
read the original abstract

Generative Adversarial Networks (GANs) have become the gold standard when it comes to learning generative models for high-dimensional distributions. Since their advent, numerous variations of GANs have been introduced in the literature, primarily focusing on utilization of novel loss functions, optimization/regularization strategies and network architectures. In this paper, we turn our attention to the generator and investigate the use of high-order polynomials as an alternative class of universal function approximators. Concretely, we propose PolyGAN, where we model the data generator by means of a high-order polynomial whose unknown parameters are naturally represented by high-order tensors. We introduce two tensor decompositions that significantly reduce the number of parameters and show how they can be efficiently implemented by hierarchical neural networks that only employ linear/convolutional blocks. We exhibit for the first time that by using our approach a GAN generator can approximate the data distribution without using any activation functions. Thorough experimental evaluation on both synthetic and real data (images and 3D point clouds) demonstrates the merits of PolyGAN against the state of the art.

Figures

Figures reproduced from arXiv: 1908.06571 by the authors.

Figure 1
Figure 1. Generated samples by an instance of the proposed PolyGAN. (a) Generated samples using [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Schematic illustration of the Coupled CP decomposition (for third order approximation). [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Schematic illustration of the Coupled nested CP decomposition (for third order approxima [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (16 more)
Figure 4
Figure 4. Figure 4: Synthesized data for learning the rx,sin xs signal. No activation functions are used in the generators. From left to right: (a) the data distribution, (b) ‘Orig’, (c) ‘Concat’, (d) PolyGAN. Notably, neither ‘Orig’ nor ‘Concat’ can learn to approximate different Taylor …
Figure 5
Figure 5. Figure 5: Synthesized data for MNIST with a single activation in the generator. From left to right: (a) [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: Conditional digit generation. Note that both ‘Orig’ and ‘Concat’ suffer from severe mode [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 2
Figure 2. Figure 2: Claim 1. Let ω “ ´ UT r2s z ¯ ˚ ´ UT r1s z ¯ ` UT r1s z. Then, the form of (7) is equal to: Gpzq “ β ` C "´ U T r3s z ¯ ˚ ω ` ω * (21) Proof. Applying Lemma 2 on (7), we obtain: Gpzq “ β ` C " U T r1s z ` ´ U T r3s z ¯ ˚ ´ U T r1s z ¯ ` ´ U T r2s z ¯ ˚ ´ U T r1s z ¯ ` …
Figure 7
Figure 7. Figure 7: PolyGAN models the data distribution accurately in contrast to the two baselines. [PITH_FULL_IMAGE:figures/full_fig_p016_7.png]
Figure 7
Figure 7. Figure 7: Synthesized data for learning the ‘astroid’ signal. No activation functions are used in the [PITH_FULL_IMAGE:figures/full_fig_p017_7.png]
Figure 8
Figure 8. Figure 8: Experiment on 3D synthetic data. From left to right: (a) the data distribution, (b) ‘Orig’, (c) [PITH_FULL_IMAGE:figures/full_fig_p017_8.png]
Figure 9
Figure 9. Figure 9: Experiment on 3D synthetic data (‘swiss roll’). From left to right: (a) the data distribution, [PITH_FULL_IMAGE:figures/full_fig_p017_9.png]
Figure 10
Figure 10. Figure 10: Synthesized data on ‘Gabriel’s Horn’. From left to right: (a) the data distribution, (b) [PITH_FULL_IMAGE:figures/full_fig_p018_10.png]
Figure 11
Figure 11. Figure 11: Image generation on faces (YaleB (Georghiades et al., 2001)) for a generator with linear [PITH_FULL_IMAGE:figures/full_fig_p018_11.png]
Figure 12
Figure 12. Figure 12: Conditional image generation on CIFAR10 for a generator with linear blocks and a single [PITH_FULL_IMAGE:figures/full_fig_p019_12.png]
Figure 13
Figure 13. Figure 13: Synthesized data for learning the rx, sinpxqs signal. No activation functions are used in the generators. From left to right: (a) the data distribution, (b) Coupled CP decomposition, (c) Coupled nested CP decomposition. (a) GT (b) Coupled CP decomposition (c) Coupled …
Figure 14
Figure 14. Figure 14: Synthesized data for learning the Astroid. Both models generate only plausible examples. [PITH_FULL_IMAGE:figures/full_fig_p025_14.png]
Figure 15
Figure 15. Figure 15: Experiment on 3D synthetic data. From left to right: (a) the data distribution, (b) Coupled [PITH_FULL_IMAGE:figures/full_fig_p025_15.png]
Figure 16
Figure 16. Figure 16: Experiment on 3D synthetic data (‘swiss roll’). From left to right: (a) the data distribution, [PITH_FULL_IMAGE:figures/full_fig_p026_16.png]
Figure 17
Figure 17. Figure 17: Comparison of the two decompositions on digit generation. [PITH_FULL_IMAGE:figures/full_fig_p026_17.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

42 extracted references · 31 canonical work pages

  1. [1]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION format.date year duplicate empty "emp...

  2. [2]

    @esa (Ref

    \@ifxundefined[1] #1\@undefined \@firstoftwo \@secondoftwo \@ifnum[1] #1 \@firstoftwo \@secondoftwo \@ifx[1] #1 \@firstoftwo \@secondoftwo [2] @ #1 \@temptokena #2 #1 @ \@temptokena \@ifclassloaded agu2001 natbib The agu2001 class already includes natbib coding, so you should not add it explicitly Type <Return> for now, but then later remove the command n...

  3. [3]

    \@lbibitem[] @bibitem@first@sw\@secondoftwo \@lbibitem[#1]#2 \@extra@b@citeb \@ifundefined br@#2\@extra@b@citeb \@namedef br@#2 \@nameuse br@#2\@extra@b@citeb \@ifundefined b@#2\@extra@b@citeb @num @parse #2 @tmp #1 NAT@b@open@#2 NAT@b@shut@#2 \@ifnum @merge>\@ne @bibitem@first@sw \@firstoftwo \@ifundefined NAT@b*@#2 \@firstoftwo @num @NAT@ctr \@secondoft...

  4. [4]

    @open @close @open @close and [1] URL: #1 \@ifundefined chapter * \@mkboth \@ifxundefined @sectionbib * \@mkboth * \@mkboth\@gobbletwo \@ifclassloaded amsart * \@ifclassloaded amsbook * \@ifxundefined @heading @heading NAT@ctr thebibliography [1] @ \@biblabel @NAT@ctr \@bibsetup #1 @NAT@ctr @ @openbib .11em \@plus.33em \@minus.07em 4000 4000 `\.\@m @bibit...

    work page

  5. [5]

    Towards principled methods for training generative adversarial networks

    Martin Arjovsky and L \'e on Bottou. Towards principled methods for training generative adversarial networks. In International Conference on Learning Representations (ICLR), 2017

    work page 2017

  6. [6]

    A convergence analysis of gradient descent for deep linear neural networks

    Sanjeev Arora, Nadav Cohen, Noah Golowich, and Wei Hu. A convergence analysis of gradient descent for deep linear neural networks. In International Conference on Learning Representations (ICLR), 2019

    work page 2019

  7. [7]

    Began: Boundary equilibrium generative adversarial networks

    David Berthelot, Thomas Schumm, and Luke Metz. Began: Boundary equilibrium generative adversarial networks. arXiv preprint arXiv:1703.10717, 2017

    arXiv 2017

  8. [8]

    Large scale gan training for high fidelity natural image synthesis

    Andrew Brock, Jeff Donahue, and Karen Simonyan. Large scale gan training for high fidelity natural image synthesis. In International Conference on Learning Representations (ICLR), 2019

    work page 2019

Show all 42 references
  1. [9]

    Generative adversarial networks: An overview

    Antonia Creswell, Tom White, Vincent Dumoulin, Kai Arulkumaran, Biswa Sengupta, and Anil A Bharath. Generative adversarial networks: An overview. IEEE Signal Processing Magazine, 35 0 (1): 0 53--65, 2018

    2018

  2. [10]

    From few to many: Illumination cone models for face recognition under variable lighting and pose

    Athinodoros S Georghiades, Peter N Belhumeur, and David J Kriegman. From few to many: Illumination cone models for face recognition under variable lighting and pose. IEEE Transactions on Pattern Analysis and Machine Intelligence (T-PAMI) , 0 (6): 0 643--660, 2001

    2001

  3. [11]

    Generative adversarial nets

    Ian Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio. Generative adversarial nets. In Advances in neural information processing systems (NIPS), 2014

    2014

  4. [12]

    Improved training of wasserstein gans

    Ishaan Gulrajani, Faruk Ahmed, Martin Arjovsky, Vincent Dumoulin, and Aaron C Courville. Improved training of wasserstein gans. In Advances in neural information processing systems (NIPS), pp.\ 5767--5777, 2017

    2017

  5. [13]

    Identity matters in deep learning

    Moritz Hardt and Tengyu Ma. Identity matters in deep learning. In International Conference on Learning Representations (ICLR), 2017

    2017

  6. [14]

    Gans trained by a two time-scale update rule converge to a local nash equilibrium

    Martin Heusel, Hubert Ramsauer, Thomas Unterthiner, Bernhard Nessler, and Sepp Hochreiter. Gans trained by a two time-scale update rule converge to a local nash equilibrium. In Advances in neural information processing systems (NIPS), pp.\ 6626--6637, 2017

    2017

  7. [15]

    Emerging convolutions for generative normalizing flows

    Emiel Hoogeboom, Rianne van den Berg, and Max Welling. Emerging convolutions for generative normalizing flows. In International Conference on Machine Learning (ICML), 2019

    2019

  8. [16]

    Arbitrary style transfer in real-time with adaptive instance normalization

    Xun Huang and Serge Belongie. Arbitrary style transfer in real-time with adaptive instance normalization. In IEEE Proceedings of International Conference on Computer Vision (ICCV) , pp.\ 1501--1510, 2017

    2017

  9. [17]

    Minimax estimation of neural net distance

    Kaiyi Ji and Yingbin Liang. Minimax estimation of neural net distance. In Advances in neural information processing systems (NIPS), pp.\ 3845--3854, 2018

    2018

  10. [18]

    Progressive growing of gans for improved quality, stability, and variation

    Tero Karras, Timo Aila, Samuli Laine, and Jaakko Lehtinen. Progressive growing of gans for improved quality, stability, and variation. In International Conference on Learning Representations (ICLR), 2018

    2018

  11. [19]

    A style-based generator architecture for generative adversarial networks

    Tero Karras, Samuli Laine, and Timo Aila. A style-based generator architecture for generative adversarial networks. In IEEE Proceedings of International Conference on Computer Vision and Pattern Recognition (CVPR) , 2019

    2019

  12. [20]

    Tensor decompositions and applications

    Tamara G Kolda and Brett W Bader. Tensor decompositions and applications. SIAM review, 51 0 (3): 0 455--500, 2009

    2009

  13. [21]

    The cifar-10 dataset

    Alex Krizhevsky, Vinod Nair, and Geoffrey Hinton. The cifar-10 dataset. online: http://www. cs. toronto. edu/kriz/cifar. html, 55, 2014

    2014

  14. [22]

    An analytic theory of generalization dynamics and transfer learning in deep linear networks

    Andrew K Lampinen and Surya Ganguli. An analytic theory of generalization dynamics and transfer learning in deep linear networks. In International Conference on Learning Representations (ICLR), 2019

    2019

  15. [23]

    Deep linear networks with arbitrary loss: All local minima are global

    Thomas Laurent and James Brecht. Deep linear networks with arbitrary loss: All local minima are global. In International Conference on Machine Learning (ICML), 2018

    2018

  16. [24]

    Gradient-based learning applied to document recognition

    Yann LeCun, L \'e on Bottou, Yoshua Bengio, Patrick Haffner, et al. Gradient-based learning applied to document recognition. Proceedings of the IEEE, 86 0 (11): 0 2278--2324, 1998

    1998

  17. [25]

    Are gans created equal? a large-scale study

    Mario Lucic, Karol Kurach, Marcin Michalski, Sylvain Gelly, and Olivier Bousquet. Are gans created equal? a large-scale study. In Advances in neural information processing systems (NIPS), pp.\ 700--709, 2018

    2018

  18. [26]

    Least squares generative adversarial networks

    Xudong Mao, Qing Li, Haoran Xie, Raymond YK Lau, Zhen Wang, and Stephen Paul Smolley. Least squares generative adversarial networks. In IEEE Proceedings of International Conference on Computer Vision (ICCV) , pp.\ 2813--2821. IEEE, 2017

    2017

  19. [27]

    Learning functions: when is deep better than shallow

    Hrushikesh Mhaskar, Qianli Liao, and Tomaso Poggio. Learning functions: when is deep better than shallow. arXiv preprint arXiv:1603.00988, 2016

    2016 arXiv

  20. [28]

    Spectral normalization for generative adversarial networks

    Takeru Miyato, Toshiki Kataoka, Masanori Koyama, and Yuichi Yoshida. Spectral normalization for generative adversarial networks. In International Conference on Learning Representations (ICLR), 2018

    2018

  21. [29]

    f-gan: Training generative neural samplers using variational divergence minimization

    Sebastian Nowozin, Botond Cseke, and Ryota Tomioka. f-gan: Training generative neural samplers using variational divergence minimization. In Advances in neural information processing systems (NIPS), pp.\ 271--279, 2016

    2016

  22. [30]

    Is generator conditioning causally related to gan performance? In International Conference on Machine Learning (ICML), 2018

    Augustus Odena, Jacob Buckman, Catherine Olsson, Tom B Brown, Christopher Olah, Colin Raffel, and Ian Goodfellow. Is generator conditioning causally related to gan performance? In International Conference on Machine Learning (ICML), 2018

    2018

  23. [31]

    Turbo-smt: Parallel coupled sparse matrix-tensor factorizations and applications

    Evangelos E Papalexakis, Tom M Mitchell, Nicholas D Sidiropoulos, Christos Faloutsos, Partha Pratim Talukdar, and Brian Murphy. Turbo-smt: Parallel coupled sparse matrix-tensor factorizations and applications. Statistical Analysis and Data Mining: The ASA Data Science Journal,...

    2016

  24. [32]

    Unsupervised representation learning with deep convolutional generative adversarial networks

    Alec Radford, Luke Metz, and Soumith Chintala. Unsupervised representation learning with deep convolutional generative adversarial networks. arXiv preprint arXiv:1511.06434, 2015

    2015 arXiv

  25. [33]

    Imagenet large scale visual recognition challenge

    Olga Russakovsky, Jia Deng, Hao Su, Jonathan Krause, Sanjeev Satheesh, Sean Ma, Zhiheng Huang, Andrej Karpathy, Aditya Khosla, Michael Bernstein, et al. Imagenet large scale visual recognition challenge. International Journal of Computer Vision (IJCV), 115 0 (3): 0 211--252, 2015

    2015

  26. [34]

    Improved techniques for training gans

    Tim Salimans, Ian Goodfellow, Wojciech Zaremba, Vicki Cheung, Alec Radford, and Xi Chen. Improved techniques for training gans. In Advances in neural information processing systems (NIPS), pp.\ 2234--2242, 2016

    2016

  27. [35]

    Exact solutions to the nonlinear dynamics of learning in deep linear neural networks

    Andrew M Saxe, James L McClelland, and Surya Ganguli. Exact solutions to the nonlinear dynamics of learning in deep linear neural networks. In International Conference on Learning Representations (ICLR), 2014

    2014

  28. [36]

    N. D. Sidiropoulos , L. De Lathauwer , X. Fu , K. Huang , E. E. Papalexakis , and C. Faloutsos . Tensor decomposition for signal processing and machine learning. IEEE Transactions on Signal Processing, 65 0 (13): 0 3551--3582, 2017

    2017

  29. [37]

    Tensor decomposition for signal processing and machine learning

    Nicholas D Sidiropoulos, Lieven De Lathauwer, Xiao Fu, Kejun Huang, Evangelos E Papalexakis, and Christos Faloutsos. Tensor decomposition for signal processing and machine learning. IEEE Transactions on Signal Processing, 65 0 (13): 0 3551--3582, 2017

    2017

  30. [38]

    The generalized weierstrass approximation theorem

    Marshall H Stone. The generalized weierstrass approximation theorem. Mathematics Magazine, 21 0 (5): 0 237--254, 1948

    1948

  31. [39]

    Going deeper with convolutions

    Christian Szegedy, Wei Liu, Yangqing Jia, Pierre Sermanet, Scott Reed, Dragomir Anguelov, Dumitru Erhan, Vincent Vanhoucke, and Andrew Rabinovich. Going deeper with convolutions. In IEEE Proceedings of International Conference on Computer Vision and Pattern Recognition (CVPR) ...

    2015

  32. [40]

    A note on the evaluation of generative models

    Lucas Theis, A \"a ron van den Oord, and Matthias Bethge. A note on the evaluation of generative models. In International Conference on Learning Representations (ICLR), 2016

    2016

  33. [41]

    Non-local neural networks

    Xiaolong Wang, Ross Girshick, Abhinav Gupta, and Kaiming He. Non-local neural networks. In IEEE Proceedings of International Conference on Computer Vision and Pattern Recognition (CVPR) , pp.\ 7794--7803, 2018

    2018

  34. [42]

    Self-attention generative adversarial networks

    Han Zhang, Ian Goodfellow, Dimitris Metaxas, and Augustus Odena. Self-attention generative adversarial networks. In International Conference on Machine Learning (ICML), 2019

    2019

Pith tools

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