Pith. sign in

REVIEW 3 major objections 6 minor 179 references

Boosting Statistic Learning with Synthetic Data from Pretrained Large Models

T0 review · 3 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read Synthetic data from pretrained diffusion models improves predictive modeling when filtered by statistical criteria.

desk verdict Interesting pipeline idea, but the paper's own results don't support its headline claim of consistent improvement. read the letter →

arxiv 2505.04992 v1 pith:T6JVAGZJ submitted 2025-05-08 stat.ML cs.LGstat.AP

Jialong Jiang , Wenkang Hu , Jian Huang , Yuling Jiao , Xu Liu This is my paper · ORCID
classification stat.MLcs.LGstat.AP
keywords dataaugmentationsyntheticStableDiffusionWassersteindistancetransferlearningtabularpredictivemodelingmodels
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper argues that synthetic data generated by a pretrained image diffusion model can improve predictive models trained on small tabular or image datasets, provided the synthetic samples are filtered before use. The proposed pipeline encodes each tabular row into a grayscale image, sends it through Stable Diffusion XL at varying strengths, decodes the generated images back into numbers, and then selects samples whose latent-space Wasserstein distance to the original data is small. Across linear, logistic, and image classification tasks, the filtered augmentation reduces prediction error relative to training on the original data alone. The paper also claims that only a limited fraction of the generated samples is useful, so improvements saturate as more synthetic data is added.

What carries the argument

The load-bearing object is the reversible mapping $M_i$ from a tabular block to a grayscale image (e.g. $e^{0.05v}$ or min-max normalization), paired with its inverse $M^{-1}$, so that Stable Diffusion XL refiner can operate on images and the result can be re-read as numbers while preserving column positions. The second component is the filtering step: for images, a VAE encodes real and generated images into a latent space and the Wasserstein-1 distance between the two latent distributions selects the generated images kept; for tabular data, transfer-learning source detection and p-value-based hypothesis tests serve the same role. Theorem 3.1's bound, using $W_1(P_{\mathrm{synth}},P_{\mathrm{real}})\le\epsilon$, Rademacher complexity, and Lipschitz loss, is what converts small Wasserstein distance into a small generalization gap.

What would settle it

Take a real tabular dataset, generate synthetic samples with the paper's pipeline, and compute the Wasserstein-1 distance between the decoded synthetic samples and held-out real samples in the original numeric space. If samples selected by the latent-space Wasserstein filter are not closer in the original space than randomly chosen generated samples, or if the selected samples fail to reduce test error relative to unfiltered augmentation, the central claim fails.

Watch

Extended reading notes

Core claim

The central claim is that a generate-then-filter loop turns a pretrained text-to-image model into a useful data augmenter for statistical prediction. The paper claims that by splitting the original data into two independent halves, encoding one half as grayscale matrices with a reversible map, generating synthetic images with the SD-XL refiner, decoding them back, and then using transfer-learning-based filtering (p-value tests for tabular data, Wasserstein distance in VAE latent space for images), the final training set yields lower prediction error than the original data. It presents Theorem 3.1 as a guarantee: if the synthetic distribution is within $\epsilon$ of the real distribution in Wasserstein-1 distance, then the real-data risk is bounded by the synthetic-data empirical risk plus $L_\ell\epsilon$, a Rademacher complexity term, and a finite-sample term. The empirical sections show consistent reductions on synthetic linear and logistic regression, Boston housing, GTEx, German credit, MNIST, CIFAR-10, ISIC, and cassava leaf data, with the improvement flattening as the generated sample size grows.

Load-bearing premise

The argument assumes that keeping the Wasserstein distance small in the VAE latent space keeps the synthetic and real distributions close in the original data space, so the bound in Theorem 3.1 actually applies to the algorithm; the paper asserts this without proving a link between the two metrics.

Editorial extensions

If this is right

  • If the claim holds, practitioners can treat a pretrained image diffusion model as a drop-in prior for tabular augmentation: encode, generate at controlled strength, filter, decode.
  • Filtering is not optional: unfiltered augmentation can hurt or add little (as in the ISIC results), so any deployment should include a distributional fidelity gate.
  • The saturation pattern implies a practical rule of thumb: augment with a modest number of synthetic samples and stop when validation error plateaus, since marginal gains vanish.
  • For image classification, Wasserstein, MMD, and TV filters give comparable gains, so the choice of metric is less critical than the act of filtering itself.
  • The theoretical bound suggests the gain is controlled by the Wasserstein gap between synthetic and real distributions, so better generation fidelity directly translates into better augmentation.

Reading between the lines

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

  • The same encode–generate–filter–decode loop could extend to other pretrained generative models (e.g., text or audio), but only if an invertible encoding exists; the paper does not test this.
  • The latent-space Wasserstein filter is a proxy for data-space fidelity; a direct test would compare filtering by latent distance against filtering by distance computed on the decoded numeric values. This is an extension not in the paper.
  • The finite-information saturation the paper observes might be quantified as an information bottleneck: the mutual information between the conditioning sample and the generated distribution caps how much new signal augmentation can add. The paper does not develop this.
  • On datasets well represented in the diffusion model's pretraining (CIFAR-100), filtering mattered little; this suggests the method's value concentrates on out-of-distribution or fine-grained domains. The paper hints at this but does not make it a general claim.
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 / 6 minor

Summary. This paper proposes a data augmentation framework that encodes tabular data as grayscale images, generates synthetic images with Stable Diffusion XL at varying diffusion strengths, decodes them back to tabular form, and filters the synthetic samples using p-value-based transfer-learning criteria (for tabular settings) or Wasserstein/MMD/TV distances in VAE latent space (for image settings). The manuscript presents a generalization bound (Theorem 3.1) under an assumed Wasserstein-1 bound between synthetic and real distributions, simulation studies for low- and high-dimensional linear and logistic models, and experiments on Boston Housing, GTEx, German Credit, MNIST, CIFAR-10/100, ISIC, and Cassava leaf disease. The stated central claim, made in the abstract, is that the framework yields consistent improvements in predictive performance across settings.

Significance. Should the central claim hold, the paper would offer a useful template for using pretrained diffusion models in statistical augmentation, and it does address an interesting question: how to select, among many generated samples, the ones that help downstream prediction. The manuscript deserves credit for evaluating multiple filtering metrics (Wasserstein, MMD, TV), for including high-dimensional transfer-learning simulations, and for an unusually candid Limitations section. The theory, however, is standard distribution-shift material, and the empirical support is the actual load-bearing contribution. On that point the paper fails to be convincing: several of its own experiments show filtering to be negligible or harmful, key tables omit the unfiltered augmentation control, and no error bars or code are provided. The result is that the headline claim of consistency is not established by the evidence in the manuscript.

major comments (3)
  1. [§4.2, §4.3, §4.6, Table 5] The paper's own results contradict the abstract's claim of 'consistent improvements.' In §4.2 (GTEx) and §4.3 (German Credit) the text states that the difference between unfiltered synthetic data and 'Ours' is negligible and that filtered results are adopted only 'for consistency.' In §4.6 the text states that filtering yields performance 'nearly identical' to unfiltered augmentation, and Table 5 shows the unfiltered 'None' row exceeding every Wasserstein-filtered row at sizes 1000, 1500, and 2000 (e.g., None 0.874 vs. Wass-40 0.851 at size 1000; None 0.888 vs. Wass-40 0.874 at size 2000). These are not isolated anomalies; they are regimes central to the claim that the filtering step is what makes augmentation effective. Since the framework's proposed novelty is generate-and-filter, the evidence that filtering helps must be consistent across the reported settings, and it is not.
  2. [§3.2, Theorem 3.1, Algorithm 2] The theorem's hypothesis is not established by the algorithm. Algorithm 2 computes the Wasserstein distance in the VAE latent space, while Theorem 3.1 requires W1(Psynth,Preal) ≤ ε for the distributions in the data space on which the loss ℓ is defined. The paper provides no Lipschitz or metric-distortion argument relating the latent-space distance to the data-space distance, so Eq. (5) does not follow from the filtering procedure described. As written, the theorem is a conditional statement about an assumed ε rather than a guarantee for the proposed method.
  3. [Tables 1–2, §4.5, Appendix A.12] The experiments do not provide the controls needed to attribute observed gains to the generate-and-filter framework. Table 1 omits the unfiltered 'Augmented' row on CIFAR-10, so the table cannot distinguish the effect of filtering from the effect of augmentation itself; Table 2 reports single accuracies (e.g., Gen=6 Augmented 45.71 vs. Wass 57.14) without error bars or repeated-run variance, and Appendix A.12 says metrics were computed over three independent runs with bootstrap confidence intervals but Table 2 does not report any such intervals. No standard augmentation baseline (e.g., random flips/crops) is included for image tasks, and SMOTE, cited in the introduction, is not compared in the tabular experiments. Since the central claim is empirical, these missing controls are load-bearing.
minor comments (6)
  1. [§4.5 and Appendix A.10] The main text says the fixed training set is 1,000 samples and later refers to merging with a 'fixed dataset of 2,500 samples'; Appendix A.10 repeats this ambiguity. Please specify whether 1,000 or 2,500 original samples are used and how the 2,500 relate to the generated set.
  2. [§4.7 and Appendix A.12] The main text states 1,257 training images for ISIC, while Appendix A.12 says the ISIC 2018 subset has 7,015 images with 1,000 training and 200 test samples. These numbers need to be reconciled.
  3. [Tables 3 and 5] The captions define 'None' inconsistently: Table 3 calls it unfiltered augmentation (mean of 100% tolerance), while Table 5 defines it as the mean of Wasserstein, TV, and MMD at 100% tolerance. Clarify whether 'None' is truly unfiltered or filtered at the 100% retention threshold.
  4. [Figure 17] Figure 17 labels the architecture as '5-class CIFAR-10 classification,' but Section 4.5 and Table 1 describe a 10-class CIFAR-10 experiment; correct the caption or the experimental description.
  5. [Reproducibility] No code, raw result tables, or complete hyperparameter specifications for Algorithm 1 (e.g., K and the validation criteria) are provided; for an empirical paper whose central claim rests on many small accuracy differences, this is a serious presentation gap.
  6. [Appendix figures] Several figure blocks in the appendix contain uninterpretable glyph sequences (e.g., '/uni00000013/...'); if these strings appear in the compiled PDF, they should be replaced with proper captions or rendered text before publication.

Circularity Check

1 steps flagged · score 4.0 of 10

The theoretical guarantee is obtained by identifying latent-space Wasserstein filtering with the theorem's image-space assumption, and the simulation protocol tunes the data-selection ratio on the test set.

  1. fitted input called prediction [Section 3.1, Algorithm 1, lines 2 and 17; Section 3.3.1]
    "Input: Source domainsS1,S2, target domainsT1,T2, test setDtest, ratio setP ={ρi}n i=1, batch size b [...] Return: ρ∗ = arg minρ∈P ¯ερ"

    Algorithm 1 selects the sampling ratio ρ* by minimizing the average prediction error ¯ερ evaluated on the independent test set Dtest, and the simulation section reports reduced prediction error for the 'Ours' method. This makes the reported improvement partly a consequence of tuning the augmentation ratio on the evaluation set, rather than a prediction made before seeing test labels. The quantity being reported as the method's error is minimized over ρ on the same Dtest used for the headline comparison, so the empirical gain is partially a fitted input presented as a prediction.

full rationale

The core empirical claim—that filtered synthetic data improve predictive performance—is not derived from a circular chain: it is tested against external baselines and publicly available datasets, and no uniqueness theorem or load-bearing self-citation forces the conclusions. However, two load-bearing steps reduce to internal choices or assumptions. First, the theoretical guarantee is presented as if latent-space Wasserstein filtering verifies the image-space Wasserstein condition required by Theorem 3.1; the paper explicitly switches to latent-space computation and never supplies a bound connecting the two metrics, so the theorem's premise is effectively assumed. Second, Algorithm 1 selects the data-sampling ratio by minimizing error on the test set, and the simulations then report improved error for the selected configuration, which is a form of test-set tuning rather than an out-of-sample prediction. These issues undermine the strength of the 'consistent improvements' claim and the applicability of the generalization bound, but they are not full definitional circularity of the central derivation. The synthetic-data improvement could in principle be real even if these protocol gaps were fixed, so a moderate circularity score is appropriate.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

The central claim rests on several unproven assumptions: the round-trip through Stable Diffusion preserves the joint distribution of tabular data, the latent-space Wasserstein distance is a faithful proxy for data-space distance, and the filtered synthetic distribution satisfies the epsilon-condition of Theorem 3.1. None of these is proved or empirically checked in the paper. Additionally, many hyperparameters (diffusion strength, guidance scale, filtering thresholds, number of generated images, transfer-learning tuning parameters) are chosen ad hoc per dataset, and some (batch size, sampling ratios in Algorithm 1) are not specified at all. These choices and assumptions are what the reader pays for upstream; the paper itself contributes the pipeline and the experiments.

free parameters (6)
  • Diffusion strength k = 0.001-1 (simulations); 0.15/0.8 (image datasets)
    Controls noise level in image-to-image generation; varied per dataset, with no sensitivity analysis.
  • Guidance scale = 7.5 or 15.0
    Chosen per experiment without justification.
  • Filtering threshold / retention rate = Top 60%-80% (images); tolerance 20-80% (Cassava)
    Selected per dataset; Tables 1-3 report multiple thresholds and favorable ones are highlighted.
  • Number of generated images per original = 2-24 (CIFAR/ISIC); 10 (CIFAR-100/Cassava)
    Chosen ad hoc; the paper scans Gen values and reports selected generations.
  • Transfer-learning tuning parameters C0 and delta0 = C0=2, delta0=2
    Set in glmtrans and hdtrd without sensitivity analysis.
  • Batch size b and sampling ratios P in Algorithm 1 = unspecified
    Not given in the paper, leaving Algorithm 1 incompletely specified.
assumptions (5)
  • domain assumption The reversible mapping M_i from tabular data to grayscale images preserves the joint distribution P_XY through the SD round-trip.
    Section 2 states the process 'preserving the structural integrity of the augmented data', but no proof or experiment validates distributional preservation.
  • domain assumption SD-XL image-to-image generation of the encoded tabular image produces samples from the input distribution in latent space.
    Section 3.2 relies on SD's pretrained prior to generate meaningful variations; never validated for tabular-encoded images.
  • domain assumption The Wasserstein distance computed in VAE latent space approximates the Wasserstein distance in the original data space.
    Section 3.2 computes W1 in latent space, but Theorem 3.1 needs it in data space; approximation error is not bounded.
  • ad hoc to paper After Wasserstein filtering, the synthetic distribution satisfies W1(Psynth,Preal) <= epsilon in data space.
    Assumed directly in Theorem 3.1; the algorithm never measures or controls the data-space Wasserstein distance.
  • domain assumption The split V1,V2 provides independent subsets, and using V2 to filter data generated from V1 maintains validity for downstream training.
    Section 2 calls this 'critical'; the selected synthetic set depends on V2, and the trained model is evaluated on a separate test set, so the independence claim is not established.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Boosting Statistic Learning with Synthetic Data from Pretrained Large Models." pith.science (2026). https://pith.science/paper/T6JVAGZJ

@misc{pith2026250504992,
  author       = {Pith},
  title        = {Pith review of: Boosting Statistic Learning with Synthetic Data from Pretrained Large Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/T6JVAGZJ}},
  note         = {Machine review of arXiv:2505.04992}
}
read the original abstract

The rapid advancement of generative models, such as Stable Diffusion, raises a key question: how can synthetic data from these models enhance predictive modeling? While they can generate vast amounts of datasets, only a subset meaningfully improves performance. We propose a novel end-to-end framework that generates and systematically filters synthetic data through domain-specific statistical methods, selectively integrating high-quality samples for effective augmentation. Our experiments demonstrate consistent improvements in predictive performance across various settings, highlighting the potential of our framework while underscoring the inherent limitations of generative models for data augmentation. Despite the ability to produce large volumes of synthetic data, the proportion that effectively improves model performance is limited.

Figures

Figures reproduced from arXiv: 2505.04992 by the authors.

Figure 1
Figure 1. Tabular Data Generation Framework Generated Generated X X X Original gen1 gen2 O [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Grayscale representation F1 of V1, generated using the transformation Mi(v) = e 0.05v . Each column corresponds to (x1, x2, x3, y) from right to left, satisfying y = 2X1 − X2 + 0.5X3 + ε. The generation process utilizes the StableDiffusionImg2ImgPipeline with the stable￾diffusion-xl-refiner-1.0 model. We set the diffusion strength parameter to range from 0.001 to 0.1 in increments of 0.001, with guidance scale fixed… view at source ↗
Figure 3
Figure 3. Density plots of the generated variables [ [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figures from the paper (14 more)
Figure 4
Figure 4. Figure 4: Prediction Error Comparison on Low-Dimensional Regression Simulation. ”Ours” [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]
Figure 5
Figure 5. Figure 5: Prediction Error Comparison on High-Dimensional Linear Regression Simulation. [PITH_FULL_IMAGE:figures/full_fig_p010_5.png]
Figure 6
Figure 6. Figure 6: Prediction error of generated data in high-dimensional generalized linear mod [PITH_FULL_IMAGE:figures/full_fig_p011_6.png]
Figure 7
Figure 7. Figure 7: Prediction Error Comparison on Boston Dataset. Due to CTGAN’s limitations [PITH_FULL_IMAGE:figures/full_fig_p012_7.png]
Figure 8
Figure 8. Figure 8: Prediction Error of Generated Data Based on GTex Data Set.On moderate dimension [PITH_FULL_IMAGE:figures/full_fig_p013_8.png]
Figure 9
Figure 9. Figure 9: Prediction error of generated data based on German Credit Data Set.Ours means [PITH_FULL_IMAGE:figures/full_fig_p013_9.png]
Figure 10
Figure 10. Figure 10: A grayscale representation F1 of V1 under high dimentional settings, Mi(v) = e 0.05v The generation process employs the Stable DiffusionImg2Img Pipeline with the stable-diffusion-xl-refiner-1.0 model. We specify the following prompt to guide the image generation: ”Hig…
Figure 11
Figure 11. Figure 11: A grayscale representation F1 of V1 under generalized high dimentional settings, Mi(v) = e 0.05v y = 0. This binary classification is consistent with the logistic regression framework, where the model predicts the probability of y = 1. To successfully detect and obtai…
Figure 12
Figure 12. Figure 12: A grayscale representation F1 of V1 of Boston House Price dataset, M is a column-wise min-max normalization The generation process employs the StableDiffusionImg2ImgPipeline with the stable￾diffusion-xl-refiner-1.0 model. We set the strength from 0.01 to 1 in steps of…
Figure 13
Figure 13. Figure 13: A grayscale representation F1 of V1 of German Credit Dataset, M is a column-wise min-max normalization To identify transferable sources, we employ the glmtrans method with a parameter setting of C0 = 2. For each iteration, we randomly select subsets of data to constru…
Figure 14
Figure 14. Figure 14: A grayscale representation F1 of V1 of GTex data set, M is a column-wise min-max normalization The generation process employs the StableDiffusionImg2ImgPipeline with the stable￾diffusion-xl-refiner-1.0 model. We set the strength from 0.01 to 1 in steps of 0.001 and gu…
Figure 15
Figure 15. Figure 15: Network architecture of SimpleCNN. Batch normalization and dropout layers are [PITH_FULL_IMAGE:figures/full_fig_p043_15.png]
Figure 16
Figure 16. Figure 16: Comparison of test accuracy between the baseline CNN model and the augmented [PITH_FULL_IMAGE:figures/full_fig_p043_16.png]
Figure 17
Figure 17. Figure 17: ResNet-20 architecture for 5-class CIFAR-10 classification. Basic blocks contain [PITH_FULL_IMAGE:figures/full_fig_p044_17.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

179 extracted references · 51 canonical work pages

  1. [1]

    How faithful is your synthetic data? sample-level metrics for evaluating and auditing generative models

    Ahmed Alaa, Boris Van Breugel, Evgeny S Saveliev, and Mihaela van der Schaar. How faithful is your synthetic data? sample-level metrics for evaluating and auditing generative models. In International Conference on Machine Learning , pages 290–306. PMLR, 2022

    2022

  2. [2]

    A systematic review of trustworthy and explainable 17 artificial intelligence in healthcare: Assessment of quality, bias risk, and data fusion

    Ahmed Shihab Albahri, Ali M Duhaim, Mohammed A Fadhel, Alhamzah Alnoor, Noor S Baqer, Laith Alzubaidi, Osamah Shihab Albahri, Abdullah Hussein Alamoodi, Jinshuai Bai, Asma Salhi, et al. A systematic review of trustworthy and explainable 17 artificial intelligence in healthcare: Assessment of quality, bias risk, and data fusion. Information Fusion, 96:156–...

    2023

  3. [3]

    Springer Science & Business Media, 2008

    Luigi Ambrosio, Nicola Gigli, and Giuseppe Savar´ e.Gradient Flows: in Metric Spaces and in the Space of Probability Measures . Springer Science & Business Media, 2008

    2008

  4. [4]

    Christophe Andrieu, Nando de Freitas, Arnaud Doucet, and Michael I. Jordan. An introduction to MCMC for machine learning. Machine Learning, 50(1):5–43, 2003

    2003

  5. [5]

    Bartlett

    Martin Anthony and Peter L. Bartlett. Neural Network Learning: Theoretical Founda- tions. Cambridge University Press, USA, 1st edition, 2009. ISBN 052111862X

    2009

  6. [6]

    Towards principled methods for training generative adversarial networks

    Martin Arjovsky and L´ eon Bottou. Towards principled methods for training generative adversarial networks. In International Conference on Learning Representations, 2017

    2017

  7. [7]

    Wasserstein generative adversar- ial networks

    Martin Arjovsky, Soumith Chintala, and L´ eon Bottou. Wasserstein generative adversar- ial networks. In International conference on machine learning , pages 214–223. PMLR, 2017

    2017

  8. [8]

    Geometrical methods in the theory of ordinary differential equations, volume 250

    Vladimir Igorevich Arnold. Geometrical methods in the theory of ordinary differential equations, volume 250. Springer Science & Business Media, 2012

    2012

Show all 179 references
  1. [9]

    Generating synthetic data in finance: opportunities, challenges and pitfalls

    Samuel A Assefa, Danial Dervovic, Mahmoud Mahfouz, Robert E Tillman, Prashant Reddy, and Manuela Veloso. Generating synthetic data in finance: opportunities, challenges and pitfalls. In Proceedings of the First ACM International Conference on AI in Finance , pages 1–8, 2020

    2020

  2. [10]

    Unsupervised domain adaptation by domain invariant projection

    Mahsa Baktashmotlagh, Mehrtash T Harandi, Brian C Lovell, and Mathieu Salzmann. Unsupervised domain adaptation by domain invariant projection. In Proceedings of the IEEE international conference on computer vision , pages 769–776, 2013

    2013

  3. [11]

    Cvae-gan: Fine- grained image generation through asymmetric training

    Jianmin Bao, Dong Chen, Fang Wen, Houqiang Li, and Gang Hua. Cvae-gan: Fine- grained image generation through asymmetric training. In Proceedings of the IEEE International Conference on Computer Vision (ICCV) , Oct 2017

    2017

  4. [12]

    Rademacher and gaussian complexities: Risk bounds and structural results

    Peter L Bartlett and Shahar Mendelson. Rademacher and gaussian complexities: Risk bounds and structural results. Journal of Machine Learning Research , 3(Nov):463–482, 2002

    2002

  5. [13]

    Bartlett, Nick Harvey, Christopher Liaw, and Abbas Mehrabian

    Peter L. Bartlett, Nick Harvey, Christopher Liaw, and Abbas Mehrabian. Nearly-tight vc-dimension and pseudodimension bounds for piecewise linear neural networks. Journal of Machine Learning Research , 20(63):1–17, 2019. URL http://jmlr.org/papers/ v20/17-612.html

    2019

  6. [14]

    Variational Algorithms for Approximate Bayesian Inference

    Matthew J Beal. Variational Algorithms for Approximate Bayesian Inference . PhD thesis, University College London, 2003

    2003

  7. [15]

    Generating synthetic data for machine learning

    Jens Becker and Lars Schmidt-Thieme. Generating synthetic data for machine learning. Data Mining and Knowledge Discovery , 32(5):1350–1376, 2018. 18

    2018

  8. [16]

    A theory of learning from different domains

    Shai Ben-David, John Blitzer, Koby Crammer, Alex Kulesza, Fernando Pereira, and Jennifer Wortman Vaughan. A theory of learning from different domains. Machine learning, 79:151–175, 2010

    2010

  9. [17]

    Eyal Betzalel, Dvir Avidan, Adam Kass, Avihai Meiri, and Roy J. J. Neufeld. Evaluation metrics for generative models: An empirical study. Machine Learning and Knowledge Extraction, 6(3):1531–1544, 2024

    2024

  10. [18]

    Deep neural networks for nonparametric interaction models with diverging dimension

    Sohom Bhattacharya, Jianqing Fan, and Debarghya Mukherjee. Deep neural networks for nonparametric interaction models with diverging dimension. arXiv e-prints, pages arXiv–2302, 2023

    2023

  11. [19]

    Mathematical statistics: basic ideas and selected topics, volumes I-II package

    Peter J Bickel and Kjell A Doksum. Mathematical statistics: basic ideas and selected topics, volumes I-II package . Chapman and Hall/CRC, 2015

    2015

  12. [20]

    The zig-zag process and super- efficient sampling for bayesian analysis of big data

    Joris Bierkens, Paul Fearnhead, Gareth Roberts, et al. The zig-zag process and super- efficient sampling for bayesian analysis of big data. The Annals of Statistics , 47(3): 1288–1320, 2019

    2019

  13. [21]

    Probability and measure

    Patrick Billingsley. Probability and measure. John Wiley & Sons, 2017

    2017

  14. [22]

    Incipient alzheimer’s disease: microarray correlation analyses reveal major transcriptional and tumor suppressor responses

    Eric M Blalock, James W Geddes, Kuey Chu Chen, Nada M Porter, William R Markes- bery, and Philip W Landfield. Incipient alzheimer’s disease: microarray correlation analyses reveal major transcriptional and tumor suppressor responses. Proceedings of the National Academy of Scie...

    2004

  15. [23]

    Blei, Alp Kucukelbir, and Jon D

    David M. Blei, Alp Kucukelbir, and Jon D. McAuliffe. Variational inference: A review for statisticians. Journal of the American Statistical Association , 112(518):859–877, 2017

    2017

  16. [24]

    eummd: efficiently computing the mmd two-sample test statistic for univariate data

    Dean A Bodenham and Yoshinobu Kawahara. eummd: efficiently computing the mmd two-sample test statistic for univariate data. Statistics and Computing , 33(5):110, 2023

    2023

  17. [25]

    A review of feature selection methods on synthetic data

    Ver´ onica Bol´ on-Canedo, Noelia S´ anchez-Maro˜ no, and Amparo Alonso-Betanzos. A review of feature selection methods on synthetic data. Knowledge and information systems, 34:483–519, 2013

    2013

  18. [26]

    The bouncy particle sampler: A nonreversible rejection-free markov chain monte carlo method

    Alexandre Bouchard-Cote, Sebastian J Vollmer, and Arnaud Doucet. The bouncy particle sampler: A nonreversible rejection-free markov chain monte carlo method. Journal of the American Statistical Association , 113(522):855–867, 2018

    2018

  19. [27]

    Skin cancer classification using convolutional neural networks: systematic review

    Titus Josef Brinker, Achim Hekler, Jochen Sven Utikal, Niels Grabe, Dirk Schaden- dorf, Joachim Klode, Carola Berking, Theresa Steeb, Alexander H Enk, and Christof Von Kalle. Skin cancer classification using convolutional neural networks: systematic review. Journal of medical ...

    2018

  20. [28]

    Jones, and Xiao-Li Meng

    Steve Brooks, Andrew Gelman, Galin L. Jones, and Xiao-Li Meng. Handbook of Markov Chain Monte Carlo . CRC Press, 2011. 19

    2011

  21. [29]

    The genotype-tissue expression (gtex) project

    Latarsha J Carithers and Helen M Moore. The genotype-tissue expression (gtex) project. Biopreservation and biobanking, 13(5):307, 2015

    2015

  22. [30]

    Chawla, Kevin W

    Nitesh V. Chawla, Kevin W. Bowyer, Lawrence O. Hall, and Wang Kegelmeyer. SMOTE: Synthetic minority over-sampling technique. Journal of Artificial Intelligence Research, 16:321–357, 2002. URL https://www.jair.org/index.php/jair/article/ view/10301

    2002

  23. [31]

    A unified particle-optimization framework for scalable bayesian sampling

    Changyou Chen, Ruiyi Zhang, Wenlin Wang, Bai Li, and Liqun Chen. A unified particle-optimization framework for scalable bayesian sampling. In UAI, 2018

    2018

  24. [32]

    Stochastic gradient Hamiltonian Monte Carlo

    Tianqi Chen, Emily Fox, and Carlos Guestrin. Stochastic gradient Hamiltonian Monte Carlo. In Eric P. Xing and Tony Jebara, editors, Proceedings of the 31st International Conference on Machine Learning , volume 32 of Proceedings of Machine Learning Research, pages 1683–1691. PM...

    2014

  25. [33]

    Boosting synthetic data generation with effective nonlinear causal discovery

    Martina Cinquini, Fosca Giannotti, and Riccardo Guidotti. Boosting synthetic data generation with effective nonlinear causal discovery. In 2021 IEEE Third International Conference on Cognitive Machine Intelligence (CogMI) , 2021

    2021

  26. [34]

    The geometry of proper scoring rules

    A Philip Dawid. The geometry of proper scoring rules. Annals of the Institute of Statistical Mathematics, 59(1):77–93, 2007

    2007

  27. [35]

    Minimax esti- mation of conditional moment models

    Nishanth Dikkala, Greg Lewis, Lester Mackey, and Vasilis Syrgkanis. Minimax esti- mation of conditional moment models. Advances in Neural Information Processing Systems, 33:12248–12262, 2020

    2020

  28. [36]

    NICE: Non-linear independent components estimation

    Laurent Dinh, David Krueger, and Yoshua Bengio. NICE: Non-linear independent components estimation. In International Conference on Learning Representations, 2015

    2015

  29. [37]

    Density estimation using Real NVP

    Laurent Dinh, Jascha Sohl-Dickstein, and Samy Bengio. Density estimation using Real NVP. In International Conference on Learning Representations, 2017

    2017

  30. [38]

    Neural mean discrepancy for efficient out-of-distribution detection

    Xin Dong, Junfeng Guo, Ang Li, Wei-Te Ting, Cong Liu, and HT Kung. Neural mean discrepancy for efficient out-of-distribution detection. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 19217–19227, 2022

    2022

  31. [39]

    High-dimensional data analysis: The curses and blessings of dimensionality

    David L Donoho et al. High-dimensional data analysis: The curses and blessings of dimensionality. AMS math challenges lecture , 1(2000):32, 2000

    2000

  32. [40]

    D. C. Dowson and B. V. Landau. The fr´ echet distance between multivariate normal distributions. Journal of Mathematical Analysis and Applications , 12(3):450–455, 1982

    1982

  33. [41]

    Kennedy, Brian J

    Simon Duane, A.D. Kennedy, Brian J. Pendleton, and Duncan Roweth. Hybrid Monte Carlo. Physics Letters B , 195(2):216–222, 1987

    1987

  34. [42]

    On the geometry of Stein variational gradient descent

    Andrew Duncan, Nikolas N¨ usken, and Lukasz Szpruch. On the geometry of Stein variational gradient descent. arXiv preprint arXiv:1912.00894 , 2019. 20

    1912 arXiv

  35. [43]

    The Hastings algorithm at fifty

    D B Dunson and J E Johndrow. The Hastings algorithm at fifty. Biometrika, 107(1): 1–23, 2019

    2019

  36. [44]

    Training gener- ative neural networks via maximum mean discrepancy optimization

    Gintare Karolina Dziugaite, Daniel M Roy, and Zoubin Ghahramani. Training gener- ative neural networks via maximum mean discrepancy optimization. arXiv preprint arXiv:1505.03906, 2015

    2015 arXiv

  37. [45]

    Stein’s paradox in statistics

    Bradley Efron and Carl Morris. Stein’s paradox in statistics. Scientific American, 236 (5):119–127, 1977

    1977

  38. [46]

    An introduction to the bootstrap

    Bradley Efron and Robert J Tibshirani. An introduction to the bootstrap. CRC press, 1994

    1994

  39. [47]

    Factor augmented sparse throughput deep relu neural networks for high dimensional regression.Journal of the American Statistical Association, (just-accepted):1–28, 2023

    Jianqing Fan and Yihong Gu. Factor augmented sparse throughput deep relu neural networks for high dimensional regression.Journal of the American Statistical Association, (just-accepted):1–28, 2023

    2023

  40. [48]

    A selective overview of deep learning

    Jianqing Fan, Cong Ma, and Yiqiao Zhong. A selective overview of deep learning. Statistical Science, 36(2):264–290, 2021

    2021

  41. [49]

    How do noise tails impact on deep relu networks? arXiv preprint arXiv:2203.10418 , 2022

    Jianqing Fan, Yihong Gu, and Wen-Xin Zhou. How do noise tails impact on deep relu networks? arXiv preprint arXiv:2203.10418 , 2022

    2022 arXiv

  42. [50]

    Deep neural networks for estimation and inference

    Max H Farrell, Tengyuan Liang, and Sanjog Misra. Deep neural networks for estimation and inference. Econometrica, 89(1):181–213, 2021

    2021

  43. [51]

    On the mathematical foundations of theoretical statistics

    Ronald A Fisher. On the mathematical foundations of theoretical statistics. Philo- sophical transactions of the Royal Society of London. Series A, containing papers of a mathematical or physical character , 222(594-604):309–368, 1922

    1922

  44. [52]

    Frigyik, S

    B.A. Frigyik, S. Srivastava, and M. R. Gupta. Functional bregman divergence and bayesian estimation of distributions. IEEE Transactions on Information Theory , 54 (11):5130–5139, 2008

    2008

  45. [53]

    An image is worth one word: Personalizing text-to-image generation using textual inversion

    Rinon Gal, Yuval Alaluf, Yuval Yifrach, Or Patashnik, Yotam Simhon, Gal Chechik, and Daniel Cohen-Or. An image is worth one word: Personalizing text-to-image generation using textual inversion. arXiv preprint arXiv:2208.01618 , 2022. URL https://arxiv.org/abs/2208.01618

    2022 arXiv

  46. [54]

    Maximum mean discrepancy test is aware of adversarial attacks

    Ruize Gao, Feng Liu, Jingfeng Zhang, Bo Han, Tongliang Liu, Gang Niu, and Masashi Sugiyama. Maximum mean discrepancy test is aware of adversarial attacks. In International Conference on Machine Learning , pages 3564–3575. PMLR, 2021

    2021

  47. [55]

    Deep generative learning via variational gradient flow

    Yuan Gao, Yuling Jiao, Yang Wang, Yao Wang, Can Yang, and Shunkang Zhang. Deep generative learning via variational gradient flow. In Kamalika Chaudhuri and Ruslan Salakhutdinov, editors, Proceedings of the 36th International Conference on Machine Learning, volume 97 of Proceed...

    2019

  48. [56]

    Deep generative learning with Euler particle transport

    Yuan Gao, Jian Huang, Yuling Jiao, Jin Liu, Xiliang Lu, and Zhijian Yang. Deep generative learning with Euler particle transport. In Proceedings of Machine Learning Research vol 145:1-33, 2021 2nd Annual Conference on Mathematical and Scientific Machine Learning, 2021

    2021

  49. [57]

    Gerber, Yanjun Han, and Yury Polyanskiy

    Patrik R. Gerber, Yanjun Han, and Yury Polyanskiy. Minimax optimal testing by classification. In Gergely Neu and Lorenzo Rosasco, editors, Proceedings of Thirty Sixth Conference on Learning Theory , volume 195 of Proceedings of Machine Learning Research, pages 5395–5432. PMLR,...

    2023

  50. [58]

    Gershman, M

    S. Gershman, M. Hoffman, and D. Blei. Nonparametric variational inference. ICML, 2012

    2012

  51. [59]

    Cambridge Series in Statistical and Probabilistic Mathematics

    Evarist Gin´ e and Richard Nickl.Mathematical Foundations of Infinite-Dimensional Statistical Models. Cambridge Series in Statistical and Probabilistic Mathematics. Cambridge University Press, 2021. doi: 10.1017/9781009022811

    2021 doi

  52. [60]

    Riemann manifold Langevin and Hamiltonian Monte Carlo methods

    Mark Girolami and Ben Calderhead. Riemann manifold Langevin and Hamiltonian Monte Carlo methods. Journal of the Royal Statistical Society: Series B (Statistical Methodology), 73(2):123–214, 2011

    2011

  53. [61]

    Strictly proper scoring rules, prediction, and estimation

    Tilmann Gneiting and Adrian E Raftery. Strictly proper scoring rules, prediction, and estimation. Journal of the American statistical Association , 102(477):359–378, 2007

    2007

  54. [62]

    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 27 , pages 2672–2680. Curran Associates, Inc., 2014

    2014

  55. [63]

    Representations of knowledge in complex systems

    Ulf Grenander and Michael I Miller. Representations of knowledge in complex systems. Journal of the Royal Statistical Society: Series B (Methodological) , 56(4):549–581, 1994

    1994

  56. [64]

    A kernel method for the two-sample-problem

    Arthur Gretton, Karsten Borgwardt, Malte Rasch, Bernhard Sch¨ olkopf, and Alex Smola. A kernel method for the two-sample-problem. Advances in neural information processing systems, 19, 2006

    2006

  57. [65]

    A kernel two-sample test

    Arthur Gretton, Karsten M Borgwardt, Malte J Rasch, Bernhard Sch¨ olkopf, and Alexander Smola. A kernel two-sample test. The Journal of Machine Learning Research, 13(1):723–773, 2012

    2012

  58. [66]

    On the (statistical) detection of adversarial examples

    Kathrin Grosse, Praveen Manoharan, Nicolas Papernot, Michael Backes, and Patrick McDaniel. On the (statistical) detection of adversarial examples. arXiv preprint arXiv:1702.06280, 2017

    2017 arXiv

  59. [67]

    Improved training of wasserstein gans

    Ishaan Gulrajani, Faruk Ahmed, Martin Arjovsky, Vincent Dumoulin, and Aaron C Courville. Improved training of wasserstein gans. Advances in neural information processing systems, 30, 2017. 22

    2017

  60. [68]

    Gutmann and J

    M. Gutmann and J. I. Hirayama. Bregman divergence as general framework to estimate unnormalized statistical models. In Conference on Uai, 2011

    2011

  61. [69]

    Gutmann and A Hyv¨ arinen

    M. Gutmann and A Hyv¨ arinen. Noise-contrastive estimation: A new estimation principle for unnormalized statistical models. Journal of Machine Learning Research, 9: 297–304, 2010

    2010

  62. [70]

    Pre-trained models: Past, present and future

    Xu Han, Zhengyan Zhang, Ning Ding, Yuxian Gu, Xiao Liu, Yuqi Huo, Jiezhong Qiu, Yuan Yao, Ao Zhang, Liang Zhang, Wentao Han, Minlie Huang, Qin Jin, Yanyan Lan, Yang Liu, Zhiyuan Liu, Zhiwu Lu, Xipeng Qiu, Ruihua Song, Jie Tang, Ji-Rong Wen, Jinhui Yuan, Wayne Xin Zhao, and Jun...

    2021 doi

  63. [71]

    Keith Hastings

    W. Keith Hastings. Monte Carlo sampling methods using Markov chains and their applications. Biometrika, 57(1):97–109, 1970

    1970

  64. [72]

    Deep residual learning for image recognition

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 770–778, 2016

    2016

  65. [73]

    Feedback-guided data synthesis for imbalanced classification

    Reyhane Askari Hemmat, Behnam Gholami, and Zohreh Azimifar. Feedback-guided data synthesis for imbalanced classification. arXiv preprint arXiv:2310.00158 , 2023. URL https://arxiv.org/abs/2310.00158

    2023 arXiv

  66. [74]

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

    Martin Heusel, Hubert Ramsauer, Thomas Unterthiner, Bernhard Nessler, Georg Friedrich Handl, Michael Widmann, Jakub Konr´ ad, Ilias Potamitis, Heinrich Herdin, Xia Hua, Asja Fischer, Seiji Takeda, Thomas Kreil, Michael H¨ oglinger, G¨ unter Klambauer, Andreas Mayr, and Sepp Ho...

    2017

  67. [75]

    Daniel Freeman, Ananya Kumar, and Yi Zhang

    Greg Hines, C. Daniel Freeman, Ananya Kumar, and Yi Zhang. Improving the scaling laws of synthetic data with deliberate practice. arXiv preprint arXiv:2502.15588 , 2025. URL https://arxiv.org/abs/2502.15588

    2025 arXiv

  68. [76]

    Hoffman and Andrew Gelman

    Matthew D. Hoffman and Andrew Gelman. The No-U-Turn Sampler: Adaptively setting path lengths in Hamiltonian Monte Carlo. Journal of Machine Learning Research , 15 (47):1593–1623, 2014

    2014

  69. [77]

    Stochastic variational inference

    Matthew D Hoffman, David M Blei, Chong Wang, and John Paisley. Stochastic variational inference. Journal of Machine Learning Research , 2013

    2013

  70. [78]

    Bayesian power steering: An effective approach for domain adaptation of diffusion models

    Ding Huang, Ting Li, and Jian Huang. Bayesian power steering: An effective approach for domain adaptation of diffusion models. arXiv preprint arXiv:2406.03683 , 2024. 23

    2024 arXiv

  71. [79]

    An error analysis of generative adversarial networks for learning distributions

    Jian Huang, Yuling Jiao, Zhen Li, Shiao Liu, Yang Wang, and Yunfei Yang. An error analysis of generative adversarial networks for learning distributions. The Journal of Machine Learning Research, 23(1):5047–5089, 2022

    2022

  72. [80]

    Evaluating aleatoric uncertainty via conditional generative models

    Ziyi Huang, Henry Lam, and Haofeng Zhang. Evaluating aleatoric uncertainty via conditional generative models. arXiv preprint arXiv:2206.04287 , 2022

    2022 arXiv

  73. [81]

    Batch normalization: Accelerating deep network training by reducing internal covariate shift

    Sergey Ioffe. Batch normalization: Accelerating deep network training by reducing internal covariate shift. arXiv preprint arXiv:1502.03167 , 2015

    2015 arXiv

  74. [82]

    Efficient statistical tests: A neural tangent kernel approach

    Sheng Jia, Ehsan Nezhadarya, Yuhuai Wu, and Jimmy Ba. Efficient statistical tests: A neural tangent kernel approach. In International Conference on Machine Learning , pages 4893–4903. PMLR, 2021

    2021

  75. [83]

    Deep nonparametric regression on approximate manifolds: Nonasymptotic error bounds with polynomial prefactors

    Yuling Jiao, Guohao Shen, Yuanyuan Lin, and Jian Huang. Deep nonparametric regression on approximate manifolds: Nonasymptotic error bounds with polynomial prefactors. The Annals of Statistics , 51(2):691–716, 2023

    2023

  76. [84]

    The variational formulation of the Fokker–Planck equation

    Richard Jordan, David Kinderlehrer, and Felix Otto. The variational formulation of the Fokker–Planck equation. SIAM Journal on Mathematical Analysis , 29(1):1–17, 1998

    1998

  77. [85]

    Foundations of modern probability, volume 2

    Olav Kallenberg and Olav Kallenberg. Foundations of modern probability, volume 2. Springer, 1997

    1997

  78. [86]

    Statistical analysis of distance estimators with density differences and density ratios

    Takafumi Kanamori and Masashi Sugiyama. Statistical analysis of distance estimators with density differences and density ratios. Entropy, 16(2):921–942, 2014

    2014

  79. [87]

    Fill-up: Balancing long-tailed data with generative models

    Utkarsh Khurana, Amogh Joshi, Abhinav Kumar, Manik Varma, and Pratik Chaud- hari. Fill-up: Balancing long-tailed data with generative models. arXiv preprint arXiv:2306.07200, 2023. URL https://arxiv.org/abs/2306.07200

    2023 arXiv

  80. [88]

    Global and local two-sample tests via regression

    Ilmun Kim, Ann B Lee, and Jing Lei. Global and local two-sample tests via regression. 2019

    2019

  81. [89]

    Classification accuracy as a proxy for two-sample testing

    Ilmun Kim, Aaditya Ramdas, Aarti Singh, and Larry Wasserman. Classification accuracy as a proxy for two-sample testing. 2021

    2021

  82. [90]

    Auto-encoding variational bayes

    Diederik P Kingma. Auto-encoding variational bayes. arXiv preprint arXiv:1312.6114 , 2013

    2013 arXiv

  83. [91]

    Auto-encoding variational bayes

    Diederik P Kingma and Max Welling. Auto-encoding variational bayes. In ICLR, 2014

    2014

  84. [92]

    A non- asymptotic analysis for stein variational gradient descent

    Anna Korba, Adil Salim, Michael Arbel, Giulia Luise, and Arthur Gretton. A non- asymptotic analysis for stein variational gradient descent. Advances in Neural Informa- tion Processing Systems, 33, 2020

    2020

  85. [93]

    Kernel stein discrepancy descent

    Anna Korba, Pierre-Cyril Aubin-Frankowski, Szymon Majewski, and Pierre Ablin. Kernel stein discrepancy descent. In International Conference on Machine Learning , pages 5719–5730. PMLR, 2021. 24

    2021

  86. [94]

    Imagenet classification with deep convolutional neural networks

    Alex Krizhevsky, Ilya Sutskever, and Geoffrey E Hinton. Imagenet classification with deep convolutional neural networks. Advances in neural information processing systems, 25, 2012

    2012

  87. [95]

    Automl two-sample test.Advances in Neural Information Processing Systems, 35:15929–15941, 2022

    Jonas M K¨ ubler, Vincent Stimper, Simon Buchholz, Krikamol Muandet, and Bernhard Sch¨ olkopf. Automl two-sample test.Advances in Neural Information Processing Systems, 35:15929–15941, 2022

    2022

  88. [96]

    On information and sufficiency

    Solomon Kullback and Richard A Leibler. On information and sufficiency. The annals of mathematical statistics , 22(1):79–86, 1951

    1951

  89. [97]

    Gradient-based learning applied to document recognition

    Yann LeCun, L´ eon Bottou, Yoshua Bengio, and Patrick Haffner. Gradient-based learning applied to document recognition. Proceedings of the IEEE, 86(11):2278–2324, 1998

    1998

  90. [98]

    Finite Difference Methods for Ordinary and Partial Differential Equations: Steady-state and Time-dependent Problems , volume 98

    Randall J LeVeque. Finite Difference Methods for Ordinary and Partial Differential Equations: Steady-state and Time-dependent Problems , volume 98. SIAM, 2007

    2007

  91. [99]

    Mmd gan: Towards deeper understanding of moment matching network

    Chun-Liang Li, Wei-Cheng Chang, Yu Cheng, Yiming Yang, and Barnab´ as P´ oczos. Mmd gan: Towards deeper understanding of moment matching network. Advances in neural information processing systems, 30, 2017

    2017

  92. [100]

    Adversarial learning of a sampler based on an unnormalized distribution

    Chunyuan Li, Ke Bai, Jianqiao Li, Guoyin Wang, Changyou Chen, and Lawrence Carin. Adversarial learning of a sampler based on an unnormalized distribution. In The 22nd International Conference on Artificial Intelligence and Statistics , pages 3302–3311, 2019

    2019

  93. [101]

    Pacgan: The power of two samples in generative adversarial networks

    Zinan Lin, Ashish Khetan, Giulia Fanti, and Sewoong Oh. Pacgan: The power of two samples in generative adversarial networks. Advances in neural information processing systems, 31, 2018

    2018

  94. [102]

    Understanding and accelerating particle-based variational inference

    Chang Liu, Jingwei Zhuo, Pengyu Cheng, Ruiyi Zhang, and Jun Zhu. Understanding and accelerating particle-based variational inference. In Kamalika Chaudhuri and Ruslan Salakhutdinov, editors, Proceedings of the 36th International Conference on Machine Learning, volume 97 of Pro...

    2019

  95. [103]

    Understanding MCMC dynamics as flows on the Wasserstein space

    Chang Liu, Jingwei Zhuo, and Jun Zhu. Understanding MCMC dynamics as flows on the Wasserstein space. In Kamalika Chaudhuri and Ruslan Salakhutdinov, editors, Proceedings of the 36th International Conference on Machine Learning , volume 97 of Proceedings of Machine Learning Res...

    2019

  96. [104]

    Learning deep kernels for non-parametric two-sample tests

    Feng Liu, Wenkai Xu, Jie Lu, Guangquan Zhang, Arthur Gretton, and Danica J Sutherland. Learning deep kernels for non-parametric two-sample tests. In International conference on machine learning, pages 6316–6326. PMLR, 2020

    2020

  97. [105]

    Stein variational gradient descent as gradient flow

    Qiang Liu. Stein variational gradient descent as gradient flow. In I. Guyon, U. V. Luxburg, S. Bengio, H. Wallach, R. Fergus, S. Vishwanathan, and R. Garnett, editors, 25 Advances in Neural Information Processing Systems , volume 30. Curran Associates, Inc., 2017

    2017

  98. [106]

    Stein variational gradient descent: A general purpose bayesian inference algorithm

    Qiang Liu and Dilin Wang. Stein variational gradient descent: A general purpose bayesian inference algorithm. In D. Lee, M. Sugiyama, U. Luxburg, I. Guyon, and R. Garnett, editors, Advances in Neural Information Processing Systems , volume 29. Curran Associates, Inc., 2016

    2016

  99. [107]

    Wasserstein generative learning of conditional distribution

    Shiao Liu, Xingyu Zhou, Yuling Jiao, and Jian Huang. Wasserstein generative learning of conditional distribution. arXiv preprint arXiv:2112.10039 , 2021

    2021 arXiv

  100. [108]

    Sliced-Wasserstein flows: Nonparametric generative modeling via optimal transport and diffusions

    Antoine Liutkus, Umut Simsekli, Szymon Majewski, Alain Durmus, and Fabian-Robert St¨ oter. Sliced-Wasserstein flows: Nonparametric generative modeling via optimal transport and diffusions. In Kamalika Chaudhuri and Ruslan Salakhutdinov, editors, Proceedings of the 36th Interna...

    2019

  101. [109]

    Revisiting classifier two-sample tests

    David Lopez-Paz and Maxime Oquab. Revisiting classifier two-sample tests. arXiv preprint arXiv:1610.06545, 2016

    2016 arXiv

  102. [110]

    Scaling limit of the Stein variational gradient descent: The mean field regime

    Jianfeng Lu, Yulong Lu, and James Nolen. Scaling limit of the Stein variational gradient descent: The mean field regime. SIAM Journal on Mathematical Analysis , 51 (2):648–671, 2019

    2019

  103. [111]

    Bootstrap and wild bootstrap for high dimensional linear models

    Enno Mammen. Bootstrap and wild bootstrap for high dimensional linear models. The annals of statistics , 21(1):255–285, 1993

    1993

  104. [112]

    A non-parametric test to detect data-copying in generative models

    Casey Meehan, Kamalika Chaudhuri, and Sanjoy Dasgupta. A non-parametric test to detect data-copying in generative models. In International Conference on Artificial Intelligence and Statistics , 2020

    2020

  105. [113]

    Equation of state calculations by fast computing machines

    Nicholas Metropolis, Arianna W Rosenbluth, Marshall N Rosenbluth, Augusta H Teller, and Edward Teller. Equation of state calculations by fast computing machines. The journal of Chemical Physics , 21(6):1087–1092, 1953

    1953

  106. [114]

    Learning in implicit generative models

    Shakir Mohamed and Balaji Lakshminarayanan. Learning in implicit generative models. arXiv preprint arXiv:1610.03483 , 2016

    2016 arXiv

  107. [115]

    Sobolev descent

    Youssef Mroueh, Tom Sercu, and Anant Raj. Sobolev descent. In Kamalika Chaudhuri and Masashi Sugiyama, editors, Proceedings of the Twenty-Second International Con- ference on Artificial Intelligence and Statistics , volume 89 of Proceedings of Machine Learning Research, pages ...

    2019

  108. [116]

    Integral probability metrics and their generating classes of functions

    Alfred M¨ uller. Integral probability metrics and their generating classes of functions. Advances in applied probability, 29(2):429–443, 1997

    1997

  109. [117]

    Reliable fidelity and diversity metrics for generative models

    Muhammad Ferjad Naeem, Seong Joon Oh, Youngjung Uh, Yunjey Choi, and Jaejun Yoo. Reliable fidelity and diversity metrics for generative models. In International Conference on Machine Learning, pages 7176–7185. PMLR, 2020. 26

    2020

  110. [118]

    Adaptive approximation and generalization of deep neural network with intrinsic dimensionality

    Ryumei Nakada and Masaaki Imaizumi. Adaptive approximation and generalization of deep neural network with intrinsic dimensionality. The Journal of Machine Learning Research, 21(1):7018–7055, 2020

    2020

  111. [119]

    Radford M. Neal. MCMC Using Hamiltonian Dynamics , chapter 5. CRC Press, 2011

    2011

  112. [120]

    Large sample estimation and hypothesis testing

    Whitney K Newey and Daniel McFadden. Large sample estimation and hypothesis testing. Handbook of econometrics, 4:2111–2245, 1994

    1994

  113. [121]

    Dataset diffusion: Diffusion- based synthetic data generation for pixel-level semantic segmentation

    Quang Nguyen, Truong Vu, Anh Tran, and Khoi Nguyen. Dataset diffusion: Diffusion- based synthetic data generation for pixel-level semantic segmentation. Advances in Neural Information Processing Systems, 36, 2024

    2024

  114. [122]

    Estimating divergence functionals and the likelihood ratio by convex risk minimization

    XuanLong Nguyen, Martin J Wainwright, and Michael I Jordan. Estimating divergence functionals and the likelihood ratio by convex risk minimization. IEEE Transactions on Information Theory , 56(11):5847–5861, 2010

    2010

  115. [123]

    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 29 (NIPS 2016) , pages 271–279. Curran Associates, Inc., October 2016

    2016

  116. [124]

    A survey on transfer learning

    Sinno Jialin Pan and Qiang Yang. A survey on transfer learning. IEEE Transactions on knowledge and data engineering , 22(10):1345–1359, 2009

    2009

  117. [125]

    Sdxl: Improving latent diffusion models for high-resolution image synthesis

    Dustin Podell, Zion English, Kyle Lacey, Andreas Blattmann, Tim Dockhorn, Jonas M¨ uller, Joe Penna, and Robin Rombach. Sdxl: Improving latent diffusion models for high-resolution image synthesis. arXiv preprint arXiv:2307.01952 , 2023

    2023 arXiv

  118. [126]

    A note on standard borel and related spaces

    Chris Preston. A note on standard borel and related spaces. Journal of Contemporary Mathematical Analysis, 44:63–71, 2009

    2009

  119. [127]

    Linear models and generalizations, 2008

    C Radhakrishna Rao. Linear models and generalizations, 2008

    2008

  120. [128]

    Conditional generative moment- matching networks

    Yong Ren, Jun Zhu, Jialian Li, and Yucen Luo. Conditional generative moment- matching networks. Advances in Neural Information Processing Systems , 29, 2016

    2016

  121. [129]

    Neural stein critics with staged l 2-regularization

    Matthew Repasky, Xiuyuan Cheng, and Yao Xie. Neural stein critics with staged l 2-regularization. IEEE Transactions on Information Theory , 2023

    2023

  122. [130]

    Variational inference with normalizing flows

    Danilo Rezende and Shakir Mohamed. Variational inference with normalizing flows. In Francis Bach and David Blei, editors, Proceedings of the 32nd International Conference on Machine Learning, volume 37 of Proceedings of Machine Learning Research, pages 1530–1538, Lille, France...

    2015

  123. [131]

    Langevin diffusions and Metropolis-Hastings algorithms

    Gareth O Roberts and Osnat Stramer. Langevin diffusions and Metropolis-Hastings algorithms. Methodology and computing in applied probability , 4(4):337–357, 2002

    2002

  124. [132]

    Roberts and Richard L

    Gareth O. Roberts and Richard L. Tweedie. Exponential convergence of Langevin distributions and their discrete approximations. Bernoulli, 2(4):341 – 363, 1996. 27

    1996

  125. [133]

    High-resolution image synthesis with latent diffusion models

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Bj¨ orn Ommer. High-resolution image synthesis with latent diffusion models. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages 10684–10695, 2022

    2022

  126. [134]

    Stabilizing training of generative adversarial networks through regularization

    Kevin Roth, Aurelien Lucchi, Sebastian Nowozin, and Thomas Hofmann. Stabilizing training of generative adversarial networks through regularization. In Advances in neural information processing systems, pages 2018–2028, 2017

    2018

  127. [135]

    Donald B. Rubin. Multiple Imputation for Nonresponse in Surveys . John Wiley & Sons, 1987

    1987

  128. [136]

    Mehdi S. M. Sajjadi, Olivier Bachem, Mario Lucic, Olivier Bousquet, and Sylvain Braun. Assessing generative models via precision and recall. In Advances in Neural Information Processing Systems 31 (NeurIPS 2018) , 2018. URL https://proceedings.neurips. cc/paper/2018/hash/80a2b...

    2018

  129. [137]

    Learning deep generative models

    Ruslan Salakhutdinov. Learning deep generative models. Annual Review of Statistics and Its Application, 2(1):361–385, 2015

    2015

  130. [138]

    The wasserstein proximal gradient algorithm

    Adil Salim, Anna Korba, and Giulia Luise. The wasserstein proximal gradient algorithm. arXiv preprint arXiv:2002.03035 , 2020

    2002 arXiv

  131. [139]

    Complexity analysis of stein variational gradient descent under talagrand’s inequality t1

    Adil Salim, Lukang Sun, and Peter Richt´ arik. Complexity analysis of stein variational gradient descent under talagrand’s inequality t1. arXiv preprint arXiv:2106.03076 , 2021

    2021 arXiv

  132. [140]

    Improved techniques for training gans

    Tim Salimans, Ian Goodfellow, Wojciech Zaremba, Vicki Cheung, Alec Radford, and Xi Chen. Improved techniques for training gans. Advances in neural information processing systems, 29, 2016

    2016

  133. [141]

    Optimal transport for applied mathematicians

    Filippo Santambrogio. Optimal transport for applied mathematicians . Springer, 2015

    2015

  134. [142]

    Mmd aggregated two-sample test

    Antonin Schrab, Ilmun Kim, M´ elisande Albert, B´ eatrice Laurent, Benjamin Guedj, and Arthur Gretton. Mmd aggregated two-sample test. arXiv preprint arXiv:2110.15073 , 2021

    2021 arXiv

  135. [143]

    Ksd aggregated goodness-of-fit test

    Antonin Schrab, Benjamin Guedj, and Arthur Gretton. Ksd aggregated goodness-of-fit test. Advances in Neural Information Processing Systems , 35:32624–32638, 2022

    2022

  136. [144]

    Boosting data analytics with synthetic volume expansion

    Xiaotong Shen, Yifei Liu, and Rex Shen. Boosting data analytics with synthetic volume expansion. In Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD ’21) , 2021

    2021

  137. [145]

    Deep network approximation charac- terized by number of neurons

    Zuowei Shen, Haizhao Yang, and Shijun Zhang. Deep network approximation charac- terized by number of neurons. arXiv preprint arXiv:1906.05497 , 2019

    1906 arXiv

  138. [146]

    Amortised map inference for image super-resolution

    Casper Kaae Snderby, Jose Caballero, Lucas Theis, Wenzhe Shi, and Ferenc Husz´ ar. Amortised map inference for image super-resolution. In International Conference on Learning Representations, 2017. 28

    2017

  139. [147]

    Wasserstein generative regression

    Shanshan Song, Tong Wang, Guohao Shen, Yuanyuan Lin, and Jian Huang. Wasserstein generative regression. arXiv preprint arXiv:2306.15163 , 2023

    2023 arXiv

  140. [148]

    Generative modeling by estimating gradients of the data distribution

    Yang Song and Stefano Ermon. Generative modeling by estimating gradients of the data distribution. In Advances in Neural Information Processing Systems , volume 32. Curran Associates, Inc., 2019

    2019

  141. [149]

    Score-based generative modeling through stochastic differential equations

    Yang Song, Jascha Sohl-Dickstein, Diederik P Kingma, Abhishek Kumar, Stefano Ermon, and Ben Poole. Score-based generative modeling through stochastic differential equations. In International Conference on Learning Representations, 2021

    2021

  142. [150]

    Dropout: a simple way to prevent neural networks from overfitting

    Nitish Srivastava, Geoffrey Hinton, Alex Krizhevsky, Ilya Sutskever, and Ruslan Salakhutdinov. Dropout: a simple way to prevent neural networks from overfitting. The journal of machine learning research , 15(1):1929–1958, 2014

    1929

  143. [151]

    Density ratio estimation in machine learning

    Masashi Sugiyama, Taiji Suzuki, and Takafumi Kanamori. Density ratio estimation in machine learning. Cambridge University Press, 2012

    2012

  144. [152]

    Density-ratio matching under the bregman divergence: a unified framework of density-ratio estimation

    Masashi Sugiyama, Taiji Suzuki, and Takafumi Kanamori. Density-ratio matching under the bregman divergence: a unified framework of density-ratio estimation. Annals of the Institute of Statistical Mathematics , 64(5):1009–1044, 2012

    2012

  145. [153]

    Generative models and model criticism via optimized maximum mean discrepancy

    Danica J Sutherland, Hsiao-Yu Tung, Heiko Strathmann, Soumyajit De, Aaditya Ramdas, Alex Smola, and Arthur Gretton. Generative models and model criticism via optimized maximum mean discrepancy. arXiv preprint arXiv:1611.04488 , 2016

    2016 arXiv

  146. [154]

    Transfer learning under high-dimensional generalized linear models

    Ye Tian and Yang Feng. Transfer learning under high-dimensional generalized linear models. Journal of the American Statistical Association , 118(544):2684–2697, 2023

    2023

  147. [155]

    Markov Chains for exploring posterior distributions

    Luke Tierney. Markov Chains for exploring posterior distributions. The Annals of Statistics, 22(4):1701–1728, 1994

    1994

  148. [156]

    Effective data augmentation with diffusion models

    Brian Trabucco, Kyle Doherty, Maria Gurinas, and Ruslan Salakhutdinov. Effective data augmentation with diffusion models. In The Twelfth International Conference on Learning Representations (ICLR 2024), 2024. URL https://openreview.net/forum? id=V2W9b3mY_l

    2024

  149. [157]

    Generative adversarial nets from a density ratio estimation perspective, 2016

    Masatoshi Uehara, Issei Sato, Masahiro Suzuki, Kotaro Nakayama, and Yutaka Matsuo. Generative adversarial nets from a density ratio estimation perspective, 2016

    2016

  150. [158]

    A. W. van der Vaart. Aymptotic statistics. Cambridge University Press, 1998

    1998

  151. [159]

    Asymptotic statistics, volume 3

    Aad W Van der Vaart. Asymptotic statistics, volume 3. Cambridge university press, 2000

    2000

  152. [160]

    van der Vaart and Jon A

    Aad W. van der Vaart and Jon A. Wellner. Weak Convergence and Empirical Processes. Springer New York, 1996. doi: 10.1007/978-1-4757-2545-2

    1996 doi

  153. [161]

    Springer Science & Business Media, 2008

    C´ edric Villani.Optimal Transport: Old and New , volume 338. Springer Science & Business Media, 2008. 29

    2008

  154. [162]

    The wasserstein distances.Optimal transport: old and new, pages 93–111, 2009

    C´ edric Villani and C´ edric Villani. The wasserstein distances.Optimal transport: old and new, pages 93–111, 2009

    2009

  155. [163]

    Springer, 2009

    C´ edric Villani et al.Optimal transport: old and new , volume 338. Springer, 2009

    2009

  156. [164]

    Wainwright and Michael I

    Martin J. Wainwright and Michael I. Jordan. Graphical models, exponential families, and variational inference. Foundations and Trends in Machine Learning , 1(1):1–305, 2008

    2008

  157. [165]

    Two-sample test with kernel projected wasserstein distance

    Jie Wang, Rui Gao, and Yao Xie. Two-sample test with kernel projected wasserstein distance. arXiv preprint arXiv:2102.06449 , 2021

    2021 arXiv

  158. [166]

    All of Statistics: A Concise Course in Statistical Inference

    Larry Wasserman. All of Statistics: A Concise Course in Statistical Inference . Springer, 2004

    2004

  159. [167]

    Bayesian learning via stochastic gradient Langevin dynamics

    Max Welling and Yee Whye Teh. Bayesian learning via stochastic gradient Langevin dynamics. In Proceedings of the 28th international conference on machine learning , ICML’11, pages 681–688. ACM, 2011

    2011

  160. [168]

    Jackknife, bootstrap and other resampling methods in regression analysis

    Chien-Fu Jeff Wu. Jackknife, bootstrap and other resampling methods in regression analysis. the Annals of Statistics , 14(4):1261–1295, 1986

    1986

  161. [169]

    CTGAN: Ef- fective training of conditional GAN for tabular data

    Lei Xu, Maria Skoularidou, Aris Antonoglou, and Mihaela van der Schaar. CTGAN: Ef- fective training of conditional GAN for tabular data. In Advances in Neural Information Processing Systems 32 (NeurIPS 2019) , 2019. URL https://proceedings.neurips. cc/paper/2019/hash/2f4fe3f97...

    2019

  162. [170]

    An empirical study on evaluation metrics of generative adversarial networks

    Qiantong Xu, Gao Huang, Yang Yuan, Chuan Guo, Yu Sun, Felix Wu, and Kilian Weinberger. An empirical study on evaluation metrics of generative adversarial networks. arXiv preprint arXiv:1806.07755 , 2018

    2018 arXiv

  163. [171]

    A survey on multi-task learning

    Yu Zhang and Qiang Yang. A survey on multi-task learning. IEEE transactions on knowledge and data engineering , 34(12):5586–5609, 2021

    2021

  164. [172]

    Comparing distributions by measuring differences that affect decision making

    Shengjia Zhao, Abhishek Sinha, Yutong He, Aidan Perreault, Jiaming Song, and Stefano Ermon. Comparing distributions by measuring differences that affect decision making. In International Conference on Learning Representations, 2021

    2021

  165. [173]

    A deep generative approach to conditional sampling

    Xingyu Zhou, Yuling Jiao, Jin Liu, and Jian Huang. A deep generative approach to conditional sampling. Journal of the American Statistical Association , 118(543): 1837–1848, 2023

    2023

  166. [174]

    Variance reduction and quasi-Newton for particle- based variational inference

    Michael Zhu, Chang Liu, and Jun Zhu. Variance reduction and quasi-Newton for particle- based variational inference. In Hal Daum´ e III and Aarti Singh, editors, Proceedings of the 37th International Conference on Machine Learning , volume 119 of Proceedings of Machine Learning...

    2020

  167. [175]

    Message passing Stein variational gradient descent

    Jingwei Zhuo, Chang Liu, Jiaxin Shi, Jun Zhu, Ning Chen, and Bo Zhang. Message passing Stein variational gradient descent. In Jennifer Dy and Andreas Krause, editors, Proceedings of the 35th International Conference on Machine Learning , volume 80 of Proceedings of Machine Lea...

    2018

  168. [176]

    (9) Given the assumption W1(Preal,P synth)≤ϵ, this implies EPrealℓ− EPsynthℓ≤Lℓϵ

    Bounding the Gap between Real and Synthetic Expectations: Since ℓ is Lℓ-Lipschitz, by the Kantorovich-Rubinstein duality [163], the difference between expected losses under Preal and Psynth is bounded by: |EPrealℓ− EPsynthℓ|≤ LℓW1(Preal,P synth). (9) Given the assumption W1(Pr...

  169. [177]

    For a class of functions F ={ℓ(h,·) :h∈H} , assuming ℓ is bounded by M, a standard uniform convergence bound

    Bounding the Gap between Synthetic and Empirical Synthetic Expectations via Uniform Convergence: The difference between the expected loss under Psynth and the empirical loss under ˆPn can be bounded using Rademacher complexity. For a class of functions F ={ℓ(h,·) :h∈H} , assum...

  170. [178]

    (10) Note that Rn(F) = Rn(ℓ◦H )

    states that, with probability at least 1 −δ over the sample{zi}∼ Psynth: EPsynthℓ− E ˆPnℓ≤ 2Rn(F) +M √ log(1/δ) 2n . (10) Note that Rn(F) = Rn(ℓ◦H )

  171. [179]

    Specifically, adding the inequality from Step 2 (EPsynthℓ≤ E ˆPnℓ+2Rn(F)+M√

    Combining Bounds to relate Real to Empirical Synthetic Expectation: By combining the bounds from Step 1 and Step 2, we relate the expected loss on real data to the empirical loss on synthetic data. Specifically, adding the inequality from Step 2 (EPsynthℓ≤ E ˆPnℓ+2Rn(F)+M√... ...

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.