Pith. sign in

REVIEW 3 major objections 4 minor 55 references

Federated Diffusion Modeling with Differential Privacy for Tabular Data Synthesis

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

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

Pith's one-line read This paper proposes DP-FedTabDiff, which merges differential privacy, federated learning, and diffusion models for synthetic tabular data; at epsilon=1 it measures a 34% drop in attack risk with 15% utility and 14% fidelity loss.

desk verdict Useful engineering integration with a clean ablation study, but the headline formal privacy guarantee is not established: reported epsilons are per-client DP-SGD budgets never composed over R=3,000 federated rounds. read the letter →

arxiv 2412.16083 v2 pith:JMGPQCW6 submitted 2024-12-20 cs.LG q-fin.ST

Timur Sattarov , Marco Schreyer , Damian Borth This is my paper · ORCID
classification cs.LGq-fin.ST
keywords differentialprivacyfederatedlearningdiffusionmodelstabulardatasynthesissyntheticprivacy-utilitytrade-offDP-SGDnon-IID
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 sets out to prove that differential privacy, federated learning, and diffusion models can be combined to synthesize tabular data with a formal privacy guarantee while keeping the data usable. The proposed framework, DP-FedTabDiff, adds per-client DP-SGD noise to a tabular diffusion model and aggregates the noisy updates with federated averaging. In experiments on four real-world mixed-type datasets, a moderate privacy budget of $\varepsilon=1$ lowers the measured risk of singling out, linkability, and inference by 34% compared with no DP, at the cost of 15% lower utility and 14% lower fidelity. The result promises a practical way for regulated institutions to share synthetic data without sharing raw records.

What carries the argument

The central mechanism is the DP-FedTabDiff training loop. Each client optimizes a FinDiff denoising diffusion probabilistic model for mixed-type tabular data, applying per-sample gradient clipping and Gaussian noise (the DP-SGD update) during $\gamma$ local steps; the server then combines the client models by Federated Averaging across $R$ communication rounds. The formal privacy claim rests on the Gaussian mechanism: the noise scale is chosen so that each local update satisfies $(\varepsilon,\delta)$-DP with $\delta = N^{-1}$, and the aggregated model is asserted to inherit that guarantee.

What would settle it

Run a full federated privacy accounting (e.g., moments accountant or RDP composition) over the 3,000 communication rounds with the paper's noise scale, clipping norm, batch size, sampling rate, and $\delta=N^{-1}$, and compare the resulting $\varepsilon$ with the claimed 10, 1, and 0.2. If the composed $\varepsilon$ is materially larger, the formal privacy guarantee stated for DP-FedTabDiff is false for the reported settings.

Watch

Extended reading notes

Core claim

The discovery is that the three components can be layered without collapse, and that a clear privacy-quality frontier emerges. At $\varepsilon=1$ the framework cuts empirical attack risk by roughly a third while degrading utility and fidelity by about 15% and 14%; lowering $\varepsilon$ to 0.2 further suppresses attack risk but visibly distorts the synthetic data's structure, while raising $\varepsilon$ to 10 approaches no-DP quality with higher re-identification risk. The paper frames this as a tunable trade-off for privacy-constrained settings rather than a single optimal operating point.

Load-bearing premise

The load-bearing premise is that the privacy loss of the per-client DP-SGD updates, measured by a single-client accountant, is also the privacy loss of the global model after $R=3{,}000$ federated averaging rounds; the paper provides no composition theorem or accountant for the federated process, so if that composition is invalid the formal $\varepsilon$-DP claim collapses.

Editorial extensions

If this is right

  • A moderate privacy budget ($\varepsilon=1$) offers a workable operating point: one-third lower measured attack risk for about 15% utility loss and 14% fidelity loss.
  • Fewer local updates ($\Gamma=100$) deliver the best privacy-quality trade-off and cut training time from 28 to 3.5 hours.
  • More federated clients improve privacy and, up to five clients, improve fidelity, with diminishing returns beyond that.
  • The four federated optimization strategies tested show nearly identical utility and fidelity, with dataset-dependent privacy differences.

Reading between the lines

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

  • The stated 'formal guarantee' is only as strong as a per-client DP-SGD account; a proper composition across the 3,000 federated rounds is not reported, so the effective global epsilon may be larger than claimed.
  • The empirical privacy score is based on simulated attacks, not a mathematical proof; it is therefore a practical risk estimate that cannot replace a federated privacy accountant if a certified guarantee is the goal.
  • A natural next experiment is to fix a total privacy budget across all federated rounds and compare the resulting utility-fidelity frontier with a centralized DP diffusion baseline; the paper does not include this comparison.
  • The same client-side DP-SGD plus federated averaging recipe could be transferred to conditional tabular diffusion or time-series data, but the noise scale and clipping norm would need recalibration for each modality.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. The paper proposes DP-FedTabDiff, a framework that combines denoising diffusion probabilistic models, federated learning, and differential privacy for synthetic tabular data generation. Each client applies Opacus-based DP-SGD locally during training, and the server aggregates the resulting model updates via Federated Averaging. The authors report experimental results on four real-world tabular datasets, measuring privacy risk with Anonymeter, utility via downstream classifier accuracy, and fidelity via distributional similarity. They claim that the aggregated global model generates ε-DP compliant synthetic data with formal privacy guarantees, and they present trade-offs across privacy budgets, numbers of clients, local update counts, and federated optimization strategies. The central formal claim is that a moderate privacy budget of ε=1 yields a 34% privacy-risk reduction with only 15% utility loss and 14% fidelity loss relative to the no-DP setting.

Significance. If the formal privacy guarantee were established, the paper would be a useful contribution to privacy-preserving synthetic tabular data generation in federated settings. The empirical framework is thoughtful: it evaluates three distinct privacy risk dimensions (singling out, linkability, inference) in addition to utility and fidelity, and it studies how DP interacts with federated hyperparameters such as local updates and client counts. However, the central technical claim — that the final global model is ε-DP compliant with ε values as reported — is not supported by the material in the manuscript. The paper also omits several DP hyperparameters required for reproducibility and provides no comparison against existing DP generative models. The empirical trade-offs are plausible and internally consistent, but the formal privacy guarantee is the load-bearing contribution, and it is not demonstrated.

major comments (3)
  1. [Section III, Eq. (6), Eq. (4); Section IV-B] The claimed privacy guarantee is not established because the paper does not compose privacy loss over the R=3,000 federated communication rounds. Equation (6) describes per-client DP-SGD with noise added to clipped gradients, and Equation (4) aggregates client updates via FedAvg. The reported ε values (e.g., ε=1) appear to be the per-client Opacus privacy budgets for local training, but each client's data influences the global model in every round in which that client participates. Basic composition would multiply the per-round ε by the number of rounds, and even advanced or Rényi composition would yield a total ε substantially larger than 1 for δ=N^{-1}. The paper provides no federated privacy accountant or moments-accounting argument across rounds. Consequently, the headline claim of an ε-DP compliant global model with ε=1 is unsupported.
  2. [Section IV-B, Differential Privacy] The manuscript does not report the Gaussian noise scale σ, the gradient clipping norm C, or the client participation schedule for the reported experiments. The text states that privacy settings are adopted from [34] and that Opacus is used, but the actual values of σ and C are absent. Without these parameters, the claimed ε values cannot be reproduced or audited, and it is impossible to verify whether the reported privacy budgets correspond to the actual noise injection. This is a load-bearing omission because the paper's central claim is a formal privacy guarantee.
  3. [Section V, RQ2 and RQ3] The experimental evaluation contains no comparison against existing differentially private generative models or centralized DP baselines. All comparisons are against the non-DP version of the authors' own framework (ε=∞). The paper claims 'optimal trade-offs between privacy budgets, client configurations, and federated optimization strategies' and 'significant improvements in privacy guarantees,' but without baselines such as PATE-GAN, DP-GAN, or a centralized DP diffusion model, these claims are not substantiated. The empirical results show internal trends, but they do not demonstrate that DP-FedTabDiff is competitive with or superior to prior work.
minor comments (4)
  1. [Equation (4)] The aggregation equation contains notation errors: the term 'λX i=1' appears to be a corrupted summation sign, and '|Di| ⊆ |D|' should be '|Di| ≤ |D|' or similar. This should be corrected for readability.
  2. [Figure 5] The heatmap labels show 'DP-' as a category (presumably denoting ε=∞), which is confusing. The label should be explicit, such as 'ε=∞ (no DP)'.
  3. [Section IV-C, Privacy] The Equation (7) aggregates three risk measures by simple averaging, but these measures have different scales and distributions (as seen in Figure 7). The aggregation scheme should be justified, or the individual components should be reported separately in the main results.
  4. [Section III, Differential Privacy] The formal DP definition in Equation (5) is standard and correctly stated, but the manuscript does not specify which DP variant (e.g., approximate DP) is used when referring to ε-DP compliance. Clarifying this would help readers interpret the claims.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the DP budget is an input, and utility/fidelity/privacy are measured outputs.

full rationale

The paper's derivation chain consists of standard building blocks: per-client DP-SGD (Eq. 6) with Opacus accounting, FedAvg aggregation (Eq. 4), and the FinDiff/FedTabDiff diffusion model. The privacy budget ε is taken as an input hyperparameter (ε ∈ {0.2, 1, 10}), and the reported privacy, utility, and fidelity scores are empirical measurements on real datasets using Anonymeter, classifier accuracy, and distributional distances. No fitted parameter is renamed as a prediction, and no definitional equality between a claimed result and its inputs appears. The authors' self-citations to FinDiff and FedTabDiff supply the architectural foundation, but the DP guarantee is attributed to Opacus and to the privacy settings of Dockhorn et al., not to those prior works. The main weakness is that the paper asserts the aggregated global model is ε-DP compliant after R=3,000 federated rounds without proving composition of the per-client DP-SGD privacy loss across rounds; this is an unsupported correctness claim, not a circular argument. There is no equation in which the derived quantity equals its input by construction, so no circular step can be exhibited.

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

No new particles, forces, dimensions, or theoretical constructs are introduced; DP-FedTabDiff is an engineering composition of existing components. The ledger instead records the unstated privacy-accounting assumptions and unreported DP hyperparameters that the central claim depends on.

free parameters (3)
  • Gaussian noise scale sigma = not reported; calibrated with Opacus to target epsilon in {0.2, 1, 10}
    Equation (6) adds noise N(0, sigma^2 I) to clipped gradients; the scale sets both the privacy guarantee and the quality degradation but its numerical value is never disclosed.
  • Gradient clipping norm C = not reported
    Each per-sample gradient is clipped to norm C in Equation (6), but the value is missing from the experimental setup, which prevents exact reproduction.
  • Number of local updates Gamma = 100
    Chosen as optimal after evaluating {10, 50, 100, 500, 1000}; this is result-dependent model selection, not a parameter fitted to a target, but it is a hand-picked operating point.
assumptions (4)
  • standard math The post-processing and composition properties of differential privacy imply that federated averaging of client DP updates yields a global model with the same privacy loss as a single client update.
    This is the background fact the paper relies on in Section III and Figure 1, but it is applied without the required composition over many communication rounds.
  • domain assumption Opacus's privacy accountant, run per client for the local optimization, is sufficient to certify the privacy of the entire federated training procedure.
    The paper does not describe a federated accountant or any composition over R=3,000 rounds; it silently equates per-client accounting with global accounting.
  • domain assumption Attack-based Anonymeter scores, measuring singling out, linkability, and inference, are an appropriate measure of the formal privacy guarantee claimed for synthetic data.
    Section IV-C defines the privacy metric as these empirical risk scores, which are not a formal DP certificate and can diverge from the true epsilon.
  • domain assumption The underlying FinDiff and FedTabDiff models generate high-fidelity tabular data in the federated setting.
    The paper inherits this from the authors' prior work [10, 13] rather than re-deriving or independently validating it beyond its own ablations.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Federated Diffusion Modeling with Differential Privacy for Tabular Data Synthesis." pith.science (2026). https://pith.science/paper/JMGPQCW6

@misc{pith2026241216083,
  author       = {Pith},
  title        = {Pith review of: Federated Diffusion Modeling with Differential Privacy for Tabular Data Synthesis},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/JMGPQCW6}},
  note         = {Machine review of arXiv:2412.16083}
}
read the original abstract

The increasing demand for privacy-preserving data analytics in various domains necessitates solutions for synthetic data generation that rigorously uphold privacy standards. We introduce the DP-FedTabDiff framework, a novel integration of Differential Privacy, Federated Learning and Denoising Diffusion Probabilistic Models designed to generate high-fidelity synthetic tabular data. This framework ensures compliance with privacy regulations while maintaining data utility. We demonstrate the effectiveness of DP-FedTabDiff on multiple real-world mixed-type tabular datasets, achieving significant improvements in privacy guarantees without compromising data quality. Our empirical evaluations reveal the optimal trade-offs between privacy budgets, client configurations, and federated optimization strategies. The results affirm the potential of DP-FedTabDiff to enable secure data sharing and analytics in highly regulated domains, paving the way for further advances in federated learning and privacy-preserving data synthesis.

Figures

Figures reproduced from arXiv: 2412.16083 by the authors.

Figure 1
Figure 1. Overview of the proposed federated diffusion modeling [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Schematic representation of the proposed [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Non-IID data distribution among (3, 5, and 10) client [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Comparative evaluation of local optimization updates [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: Heatmaps illustrating the impact of DP budgets [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]
Figure 6
Figure 6. Figure 6: t-SNE visualization of (a) dataset DA, (b) synthetic data without DP, and (c-e) with DP using federated optimization across 5 clients (λ=5). As privacy increases (i.e., lower ε), DP noise degrades the structural integrity of the synthetic data. A B C D Dataset 0.00 0.0…
Figure 7
Figure 7. Figure 7: Singling Out, Linkability, and Inference risk evaluation [PITH_FULL_IMAGE:figures/full_fig_p006_7.png]
Figure 8
Figure 8. Figure 8: Privacy, utility, and fidelity evaluation using a varying [PITH_FULL_IMAGE:figures/full_fig_p007_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

55 extracted references · 34 canonical work pages

  1. [34]

    Differentially private diffusion models,

    T. Dockhorn, T. Cao, A. Vahdat, and K. Kreis, “Differentially private diffusion models,” arXiv preprint arXiv:2210.09929 , 2022

    arXiv 2022

  2. [1]

    Model Inversion Attacks that Exploit Confidence Information and Basic Countermeasures,

    M. Fredrikson, S. Jha, and T. Ristenpart, “Model Inversion Attacks that Exploit Confidence Information and Basic Countermeasures,” in Proceedings of the 22nd ACM SIGSAC conference on computer and communications security, 2015, pp. 1322–1333

    work page 2015

  3. [2]

    ML-Leaks: Model and Data Independent Membership Inference At- tacks and Defenses on Machine Learning Models,

    A. Salem, Y . Zhang, M. Humbert, P. Berrang, M. Fritz, and M. Backes, “ML-Leaks: Model and Data Independent Membership Inference At- tacks and Defenses on Machine Learning Models,” arXiv preprint arXiv:1806.01246, 2018

    arXiv 2018

  4. [3]

    On the Dangers of Stochastic Parrots: Can Language Models Be Too Big?

    E. M. Bender, T. Gebru, A. McMillan-Major, and S. Shmitchell, “On the Dangers of Stochastic Parrots: Can Language Models Be Too Big?” in Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency, 2021, pp. 610–623

    work page 2021

  5. [4]

    Advances and Open Problems in Federated Learning,

    P. Kairouz, H. B. McMahan, B. Avent, A. Bellet, M. Bennis, A. N. Bhagoji, K. Bonawitz, Z. Charles, G. Cormode, R. Cummings et al. , “Advances and Open Problems in Federated Learning,” arXiv preprint arXiv:1912.04977, 2019

    arXiv 1912

  6. [5]

    Communication-Efficient Learning of Deep Networks from Decentral- ized Data,

    B. McMahan, E. Moore, D. Ramage, S. Hampson, and B. A. y Arcas, “Communication-Efficient Learning of Deep Networks from Decentral- ized Data,” in Artificial Intelligence and Statistics . PMLR, 2017

    work page 2017

  7. [6]

    Federated Learning: Collaborative Ma- chine Learning Without Centralized Training Data,

    B. McMahan and D. Ramage, “Federated Learning: Collaborative Ma- chine Learning Without Centralized Training Data,” Google Research Blog, vol. 3, 2017

    work page 2017

  8. [7]

    Our data, ourselves: Privacy via distributed noise generation,

    C. Dwork, K. Kenthapadi, F. McSherry, I. Mironov, and M. Naor, “Our data, ourselves: Privacy via distributed noise generation,” in Advances in Cryptology-EUROCRYPT 2006. Proceedings 25 . Springer, 2006

    work page 2006

Show all 55 references
  1. [8]

    Synthetic data generation for fraud detection using gans,

    C. Charitou, S. Dragicevic, and A. d’Avila Garcez, “Synthetic data generation for fraud detection using gans,” 2021

    2021

  2. [9]

    Synthesizing test data for fraud detection systems,

    E. Barse, H. Kvarnstrom, and E. Jonsson, “Synthesizing test data for fraud detection systems,” in19th Annual Computer Security Applications Conference, 2003. Proceedings., 2003, pp. 384–394

    2003

  3. [10]

    Fedtabdiff: Federated learning of diffusion probabilistic models for synthetic mixed-type tabular data generation,

    T. Sattarov, M. Schreyer, and D. Borth, “Fedtabdiff: Federated learning of diffusion probabilistic models for synthetic mixed-type tabular data generation,” arXiv preprint arXiv:2401.06263 , 2024

    2024 arXiv

  4. [11]

    Diffusion Models Beat GANs on Image Synthesis,

    P. Dhariwal and A. Nichol, “Diffusion Models Beat GANs on Image Synthesis,” Advances in Neural Information Processing Systems , 2021

    2021

  5. [12]

    High-Resolution Image Synthesis with Latent Diffusion Models,

    R. Rombach, A. Blattmann, D. Lorenz, P. Esser, and B. Ommer, “High-Resolution Image Synthesis with Latent Diffusion Models,” in IEEE/CVF conference on computer vision and pattern recognition, 2022

    2022

  6. [13]

    Findiff: Diffusion models for financial tabular data generation,

    T. Sattarov, M. Schreyer, and D. Borth, “Findiff: Diffusion models for financial tabular data generation,” in Proceedings of the Fourth ACM International Conference on AI in Finance , 2023, pp. 64–72

    2023

  7. [14]

    A Survey on Generative Diffusion Model,

    H. Cao, C. Tan, Z. Gao, Y . Xu, G. Chen, P.-A. Heng, and S. Z. Li, “A Survey on Generative Diffusion Model,” arXiv preprint arXiv:2209.02646, 2022

    2022 arXiv

  8. [15]

    Diffusion Models: A Com- prehensive Survey of Methods and Applications,

    L. Yang, Z. Zhang, Y . Song, S. Hong, R. Xu, Y . Zhao, Y . Shao, W. Zhang, B. Cui, and M.-H. Yang, “Diffusion Models: A Com- prehensive Survey of Methods and Applications,” arXiv preprint arXiv:2209.00796, 2022

    2022

  9. [16]

    Diffusion Models in Vision: A Survey,

    F.-A. Croitoru, V . Hondru, R. T. Ionescu, and M. Shah, “Diffusion Models in Vision: A Survey,” IEEE Transactions on Pattern Analysis and Machine Intelligence , 2023

    2023

  10. [17]

    Federated Learn- ing: A Survey on Enabling Technologies, Protocols, and Applications,

    M. Aledhari, R. Razzak, R. M. Parizi, and F. Saeed, “Federated Learn- ing: A Survey on Enabling Technologies, Protocols, and Applications,” IEEE Access, vol. 8, pp. 140 699–140 725, 2020

    2020

  11. [18]

    A survey on federated learning systems: Vision, hype and reality for data privacy and protection,

    Q. Li, Z. Wen, Z. Wu, S. Hu, N. Wang, Y . Li, X. Liu, and B. He, “A survey on federated learning systems: Vision, hype and reality for data privacy and protection,” IEEE Transactions on Knowledge and Data Engineering, 2021

    2021

  12. [19]

    A Survey on Federated Learning,

    C. Zhang, Y . Xie, H. Bai, B. Yu, W. Li, and Y . Gao, “A Survey on Federated Learning,” Knowledge-Based Systems, vol. 216, 2021

    2021

  13. [20]

    Modeling tabular data using conditional gan,

    L. Xu, M. Skoularidou, A. Cuesta-Infante, and K. Veeramachaneni, “Modeling tabular data using conditional gan,” NeurIPS, vol. 32, 2019

    2019

  14. [21]

    Conditional Wasserstein GAN-based oversampling of tabular data for imbalanced learning,

    J. Engelmann and S. Lessmann, “Conditional Wasserstein GAN-based oversampling of tabular data for imbalanced learning,” Expert Systems with Applications, vol. 174, p. 114582, 2021

    2021

  15. [22]

    PATE-GAN: Generating Synthetic Data with Differential Privacy Guarantees,

    J. Jordon, J. Yoon, and M. Van Der Schaar, “PATE-GAN: Generating Synthetic Data with Differential Privacy Guarantees,” in ICLR, 2018

    2018

  16. [23]

    Differentially Private Synthetic Medical Data Generation Using Convolutional GANs,

    A. Torfi, E. A. Fox, and C. K. Reddy, “Differentially Private Synthetic Medical Data Generation Using Convolutional GANs,” Information Sciences, vol. 586, pp. 485–500, 2022

    2022

  17. [24]

    Generating high- fidelity, synthetic time series datasets with doppelganger,

    Z. Lin, A. Jain, C. Wang, G. Fanti, and V . Sekar, “Generating high- fidelity, synthetic time series datasets with doppelganger,” arXiv preprint arXiv:1909.13403, 2019

    1909 arXiv

  18. [25]

    On the privacy properties of gan- generated samples,

    Z. Lin, V . Sekar, and G. Fanti, “On the privacy properties of gan- generated samples,” in International Conference on Artificial Intelli- gence and Statistics . PMLR, 2021, pp. 1522–1530

    2021

  19. [26]

    CTAB-GAN: Effective Table Data Synthesizing,

    Z. Zhao, A. Kunar, R. Birke, and L. Y . Chen, “CTAB-GAN: Effective Table Data Synthesizing,” in Asian Conference on Machine Learning . PMLR, 2021, pp. 97–112

    2021

  20. [27]

    Tabddpm: Modelling tabular data with diffusion models,

    A. Kotelnikov, D. Baranchuk, I. Rubachev, and A. Babenko, “Tabddpm: Modelling tabular data with diffusion models,” 2022

    2022

  21. [28]

    Argmax flows and multinomial diffusion: Learning categorical distributions,

    E. Hoogeboom, D. Nielsen, P. Jaini, P. Forr ´e, and M. Welling, “Argmax flows and multinomial diffusion: Learning categorical distributions,” Advances in Neural Information Processing Systems , vol. 34, 2021

    2021

  22. [29]

    Imb-findiff: Conditional diffusion models for class imbalance synthesis of financial tabular data,

    M. Schreyer, T. Sattarov, A. Sim, and K. Wu, “Imb-findiff: Conditional diffusion models for class imbalance synthesis of financial tabular data,” in Proceedings of the 5th ACM International Conference on AI in Finance, 2024, pp. 617–625

    2024

  23. [30]

    Frauddiffuse: Diffusion-aided syn- thetic fraud augmentation for improved fraud detection,

    R. Roy, D. Tiwari, and A. Pandey, “Frauddiffuse: Diffusion-aided syn- thetic fraud augmentation for improved fraud detection,” in Proceedings of the 5th ACM International Conference on AI in Finance , 2024

    2024

  24. [31]

    Training Diffusion Models with Federated Learning: A Communication-Efficient Model for Cross-Silo Federated Image Gener- ation,

    M. de Goede, “Training Diffusion Models with Federated Learning: A Communication-Efficient Model for Cross-Silo Federated Image Gener- ation,” 2023

    2023

  25. [32]

    Phoenix: A Federated Generative Diffusion Model,

    F. V . S. Jothiraj and A. Mashhadi, “Phoenix: A Federated Generative Diffusion Model,” arXiv preprint arXiv:2306.04098 , 2023

    2023

  26. [33]

    Deep learning with differential privacy,

    M. Abadi, A. Chu, I. Goodfellow, H. B. McMahan, I. Mironov, K. Talwar, and L. Zhang, “Deep learning with differential privacy,” in Proceedings of the 2016 ACM SIGSAC conference on computer and communications security, 2016, pp. 308–318

    2016

  27. [35]

    A survey of differentially private generative adversarial net- works,

    L. Fan, “A survey of differentially private generative adversarial net- works,” in The AAAI Workshop on Privacy-Preserving Artificial Intelli- gence, vol. 8, 2020

    2020

  28. [36]

    A systematic review of federated generative models,

    A. V . Gargary and E. De Cristofaro, “A systematic review of federated generative models,” arXiv preprint arXiv:2405.16682 , 2024

    2024 arXiv

  29. [37]

    Gs-wgan: A gradient-sanitized approach for learning differentially private generators,

    D. Chen, T. Orekondy, and M. Fritz, “Gs-wgan: A gradient-sanitized approach for learning differentially private generators,” Advances in Neural Information Processing Systems , pp. 12 673–12 684, 2020

    2020

  30. [38]

    Sgde: Secure generative data exchange for cross-silo federated learning,

    E. Lomurno, A. Archetti, L. Cazzella, S. Samele, L. Di Perna, and M. Matteucci, “Sgde: Secure generative data exchange for cross-silo federated learning,” in Proceedings of the 2022 5th International Con- ference on Artificial Intelligence and Pattern Recognition , 2022

    2022

  31. [39]

    Generative models for effective ml on private, decen- tralized datasets,

    S. Augenstein, H. B. McMahan, D. Ramage, S. Ramaswamy, P. Kairouz, M. Chen et al., “Generative models for effective ml on private, decen- tralized datasets,” arXiv preprint arXiv:1911.06679 , 2019

    1911 arXiv

  32. [40]

    Feddpgan: Fed- erated differentially private generative adversarial networks framework for the detection of covid-19 pneumonia,

    L. Zhang, B. Shen, A. Barnawi, S. Xi, and N. Kumar, “Feddpgan: Fed- erated differentially private generative adversarial networks framework for the detection of covid-19 pneumonia,” Journal of Computer Security

  33. [41]

    Differentially private secure multi- party computation for federated learning in financial applications,

    D. Byrd and A. Polychroniadou, “Differentially private secure multi- party computation for federated learning in financial applications,” in Proceedings of the First ACM ICIAF , 2020, pp. 1–9

    2020

  34. [42]

    Federated and Privacy- Preserving Learning of Accounting Data in Financial Statement Audits,

    M. Schreyer, T. Sattarov, and D. Borth, “Federated and Privacy- Preserving Learning of Accounting Data in Financial Statement Audits,” in Proceedings of the Third ACM ICAIF , 2022, pp. 105–113

    2022

  35. [43]

    Deep Unsupervised Learning Using Nonequilibrium Thermodynamics,

    J. Sohl-Dickstein, E. Weiss, N. Maheswaranathan, and S. Ganguli, “Deep Unsupervised Learning Using Nonequilibrium Thermodynamics,” in International conference on machine learning . PMLR, 2015

    2015

  36. [44]

    Denoising Diffusion Probabilistic Models,

    J. Ho, A. Jain, and P. Abbeel, “Denoising Diffusion Probabilistic Models,” NeurIPS, vol. 33, pp. 6840–6851, 2020

    2020

  37. [45]

    The algorithmic foundations of differential privacy,

    C. Dwork, A. Roth et al. , “The algorithmic foundations of differential privacy,” Foundations and Trends® in Theoretical Computer Science , vol. 9, no. 3–4, pp. 211–407, 2014

    2014

  38. [46]

    Pytorch: An imperative style, high-performance deep learning library,

    A. Paszke, S. Gross, F. Massa, A. Lerer, J. Bradbury, G. Chanan, T. Killeen, Z. Lin, N. Gimelshein, L. Antiga et al. , “Pytorch: An imperative style, high-performance deep learning library,” Advances in neural information processing systems , vol. 32, 2019

    2019

  39. [47]

    Adam: A method for stochastic optimization,

    D. P. Kingma and J. Ba, “Adam: A method for stochastic optimization,” arXiv preprint arXiv:1412.6980 , 2014

    2014 arXiv

  40. [48]

    Flower: A friendly federated learning research framework,

    D. J. Beutel, T. Topal, A. Mathur, X. Qiu, J. Fernandez-Marques, Y . Gao, L. Sani, K. H. Li, T. Parcollet, P. P. B. de Gusm ˜ao, and N. D. Lane, “Flower: A friendly federated learning research framework,” 2022

    2022

  41. [49]

    Adaptive federated optimization,

    S. Reddi, Z. Charles, M. Zaheer, Z. Garrett, K. Rush, J. Kone ˇcn`y, S. Kumar, and H. B. McMahan, “Adaptive federated optimization,”arXiv preprint arXiv:2003.00295, 2020

    2003 arXiv

  42. [50]

    Federated optimization in heterogeneous networks,

    T. Li, A. K. Sahu, M. Zaheer, M. Sanjabi, A. Talwalkar, and V . Smith, “Federated optimization in heterogeneous networks,” Proceedings of Machine learning and systems , vol. 2, pp. 429–450, 2020

    2020

  43. [51]

    Opacus: User-friendly differential privacy library in PyTorch,

    A. Yousefpour, I. Shilov, A. Sablayrolles, D. Testuggine, K. Prasad, M. Malek, J. Nguyen, S. Ghosh, A. Bharadwaj, J. Zhao, G. Cormode, and I. Mironov, “Opacus: User-friendly differential privacy library in PyTorch,” arXiv preprint arXiv:2109.12298 , 2021

    2021 arXiv

  44. [52]

    A unified framework for quantifying privacy risk in synthetic data,

    M. Giomi, F. Boenisch, C. Wehmeyer, and B. Tasn ´adi, “A unified framework for quantifying privacy risk in synthetic data,” arXiv preprint arXiv:2211.10459, 2022

    2022 arXiv

  45. [53]

    Opinion 05/2014 on anonymisation techniques,

    Article 29 Data Protection Working Party, “Opinion 05/2014 on anonymisation techniques,” Tech. Rep., April 2014. [On- line]. Available: https://ec.europa.eu/justice/article-29/documentation/ opinion-recommendation/files/2014/wp216 en.pdf

    2014

  46. [54]

    Zychlinski, “dython,” 2018

    S. Zychlinski, “dython,” 2018. [Online]. Available: https://github.com/ shakedzy/dython

    2018

  47. [55]

    Scaffold: Stochastic controlled averaging for on-device federated learning,

    S. P. Karimireddy, S. Kale, M. Mohri, S. J. Reddi, S. U. Stich, and A. T. Suresh, “Scaffold: Stochastic controlled averaging for on-device federated learning,” preprint arXiv:1910.06378, 2019

    1910 arXiv

Pith tools

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