Pith. sign in

REVIEW 4 major objections 6 minor 74 references

Task Diversity in Bayesian Federated Learning: Simultaneous Processing of Classification and Regression

T0 review · 4 major / 6 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 argues that federated learning can handle correlated classification and regression tasks on the same device by coupling them through a multi-output Gaussian process, and that the resulting Bayesian method outperforms…

desk verdict A clean combination of Pólya-Gamma augmentation with multi-output GPs for mixed-task federated learning, but the uncertainty claims outrun the evidence, mainly because the sparse approximation is an unvalidated DTC projection. read the letter →

arxiv 2412.10897 v2 pith:AIRVWDUO submitted 2024-12-14 cs.LG stat.ML

Junliang Lyu , Yixuan Zhang , Xiaoling Lu , Feng Zhou This is my paper · ORCID
classification cs.LGstat.ML
keywords Bayesianfederatedlearningmulti-taskmulti-outputGaussianprocessPólya-Gammaaugmentationmean-fieldvariationalinferenceuncertaintycalibrationout-of-distributiondetectionpersonalized
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

Federated learning usually assumes every device is solving the same kind of task, but real devices often collect both categorical labels and continuous targets at once. This paper proposes pFed-Mul, a Bayesian federated method where each client models correlated classification and regression tasks jointly through a multi-output Gaussian process, and the server updates a global prior by averaging the clients' variational posteriors. The paper claims that, in few-shot settings on synthetic data, CelebA, and Dogcat, this joint modeling improves predictive accuracy, out-of-distribution detection, uncertainty calibration, and convergence rate compared with single-task federated baselines. The technical core is making classification inference tractable: Pólya-Gamma augmentation turns the non-conjugate logistic likelihood into a conditionally conjugate one, so local updates are closed-form mean-field iterations. A sympathetic reader would care because this is a concrete route to serving heterogeneous real-world clients with limited data while still getting calibrated uncertainty.

What carries the argument

The machinery is the pair formed by a multi-output Gaussian process prior and Pólya-Gamma augmentation. The MOGP prior, built from $B$ independent basis Gaussian processes with a mixing weight matrix $\mathbf{W}$, defines a cross-covariance between every pair of classification and regression latent functions, which is what lets one task type borrow strength from the other. Pólya-Gamma augmentation introduces auxiliary variables $\boldsymbol{\omega}$ that make the logistic classification likelihood Gaussian in the latent functions given $\boldsymbol{\omega}$, so the joint model becomes conditionally conjugate. Mean-field variational inference then alternates between closed-form updates: $q_1(\boldsymbol{\omega})$ is a Pólya-Gamma distribution and $q_2(\mathbf{f}) = \mathcal{N}(\mathbf{m}, \boldsymbol{\Sigma})$ with $\boldsymbol{\Sigma} = (\mathbf{H} + \mathbf{K}^{-1})^{-1}$ and $\mathbf{m} = \boldsymbol{\Sigma}\mathbf{H}\mathbf{v}$. This closed-form structure is what lets the server aggregate posteriors and maximize an averaged ELBO analytically, and it is also what the paper credits for fast convergence, via the equivalence between mean-field VI on a conditionally conjugate model and natural-gradient descent with unit step size.

What would settle it

Run pFed-Mul on the same few-shot CelebA and Dogcat splits while increasing the number of mean-field iterations and local updates, or replacing the deterministic inducing-point projection with a proper variational bound, and compare predictive accuracy, ECE, and OOD variance; if any of these change materially, the cheap approximation is responsible for the claims.

Watch

Extended reading notes

Core claim

The central claim, stated on the paper's own terms, is that task diversity on federated clients should be modeled, not ignored. pFed-Mul assumes each client has $T_r$ regression and $T_c$ classification tasks whose latent functions are drawn from a multi-output Gaussian process with a linear model of coregionalization, so correlated tasks share statistical strength. On the server, the averaged evidence lower bound over clients is maximized to update the global MOGP prior, which is then redistributed as the next round's prior. The paper reports that pFed-Mul consistently outperforms existing methods across almost all of its few-shot scenarios, with the largest gains of $0.155$ in regression mean squared error and $3.86\%$ in classification accuracy, the best expected calibration error among baselines, clearly elevated predictive variance on OOD images, and faster early-round convergence.

Load-bearing premise

The load-bearing premise is that two mean-field iterations and two local updates per round, together with the deterministic inducing-point projection, produce a posterior that is faithful enough that the reported accuracy, calibration, and OOD gains reflect the model rather than artifacts of the approximation.

Editorial extensions

If this is right

  • If pFed-Mul is correct, a federated system can train a device's correlated classification and regression tasks simultaneously, improving both relative to separate single-task models, especially with few-shot data.
  • Predictive variance from the MOGP posterior is usable as an OOD signal: masked or semantically shifted inputs show higher variance in the reported experiments.
  • Uncertainty calibration improves: pFed-Mul records the lowest expected calibration error among the compared methods on the CelebA classification task.
  • Convergence in the first ten communication rounds is faster and more stable than the compared baselines, with only two mean-field iterations and two local updates per round.
  • The global MOGP prior learned by the server can be distributed to new clients, giving them a personalized posterior after local inference with their own data.

Reading between the lines

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

  • Editorial extension: the Pólya-Gamma route the paper develops for binary classification can in principle be carried to multi-class classification through logistic-softmax or one-vs-each softmax approximations, which the paper mentions in an appendix but does not implement; testing that extension is a direct next step.
  • Editorial extension: because the empirical claims rest on only two mean-field iterations and two local updates per round, a natural stress test is to increase those counts and check whether accuracy and calibration stay stable; the paper does not report such a sensitivity curve.
  • Editorial extension: the deterministic inducing-point projection in Appendix C replaces a conditional distribution by its mean, so the reported OOD variance could understate posterior uncertainty; comparing against a full variational treatment of the inducing points would indicate how much of the calibration gain is due to the model versus the approximation.
  • Editorial extension: the framework treats task diversity as a resource for knowledge transfer rather than merely a source of heterogeneity, suggesting that future federated systems could actively schedule which task types a device contributes.
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

4 major / 6 minor

Summary. The paper proposes pFed-Mul, a personalized Bayesian federated learning method in which each client fits a multi-output Gaussian process (MOGP) over correlated classification and regression tasks, using Pólya-Gamma augmentation and mean-field variational inference to obtain closed-form local posterior updates. The server aggregates local ELBOs to update the global MOGP prior hyperparameters, including the deep-kernel backbone, and redistributes the updated prior to clients. Experiments on synthetic data, CelebA, and Dogcat claim superior predictive accuracy, calibration, OOD detection, and convergence rate. The main technical contribution is the analytical local inference scheme and a server-side ELBO optimization for the global prior.

Significance. If the claims held, the paper would make a useful step: a Bayesian FL method that handles task diversity with closed-form updates and calibrated uncertainty is valuable, and the variational derivation in Appendices A-D is internally consistent and standard. The public code, explicit analytical expressions, and the use of Pólya-Gamma augmentation to enable mean-field VI are strengths. However, the current experimental evidence does not support several of the paper's central claims, and the sparse-GP approximation used for the uncertainty quantities is not validated. These issues are substantive but fixable within the manuscript's scope.

major comments (4)
  1. [Section 5.2.2, Table 1] The statement that 'pFed-Mul consistently outperforms existing methods across almost all scenarios' is directly contradicted by the reported regression results on CelebA. For the 10-shot/20-client, 20-shot/15-client, and 50-shot/10-client settings, pFed-Mul has MSE 0.488, 0.476, and 0.301, whereas FedPer has 0.369, 0.328, and 0.261, and the single-task variant pFed-St has 0.690, 0.321, and 0.221. Thus pFed-Mul is worse than both baselines in all three settings, and the multi-task framework does not improve regression over the single-task variant on this dataset. The abstract and the 'most significant improvements' sentence need to be revised, or the result needs to be explained and supported by additional experiments.
  2. [Appendix C, Eq. (13)] Replacing the conditional distribution p(f_{x_n} | f_{x_m}) with its conditional mean is the deterministic training conditional (DTC / projected-process) approximation. This projection removes the K_{xx} - Q_{xx} term from the predictive variance and is known to underestimate posterior uncertainty when the number of inducing points M is much smaller than N. The calibration and OOD results in Section 5.3 (ECE 0.054 and the variance spikes in Figure 4) are precisely the quantities this approximation can corrupt, and no comparison is made against an exact MOGP posterior, a fully variational sparse GP, or MCMC. In addition, the experimental protocol does not state the number of inducing points M or whether the sparse approximation in Eq. (13) is active in the reported results. Please add an approximation-error study or substantially soften the uncertainty claims.
  3. [Section 5.3.2, Figure 4] The claimed 'superior' OOD detection is not supported by the reported evidence. The only result is a visual inspection of predictive variance under two masked images for pFed-Mul alone; no quantitative OOD metric (e.g., AUROC or similar) and no baseline comparison are provided. Section 5.3 states that the method achieves superior performance to previous baselines in OOD detection, but no baseline OOD numbers appear anywhere. Please add a quantitative OOD evaluation or revise the claim.
  4. [Section 5.4, Figure 5] The convergence-rate claim rests on a single trajectory over the first 10 communication rounds, with no repetitions, error bars, or significance analysis. Since all methods are run with two local updates per round, the plot does not show whether pFed-Mul's advantage comes from the Pólya-Gamma natural-gradient property or from initialization and hyperparameter choices. Please report means and standard deviations over multiple seeds and, ideally, a matched-compute comparison.
minor comments (6)
  1. [Section 5.1.1] For the synthetic experiment, all hyperparameters are initialized as the ground truth; this makes the result a sanity check of the inference machinery rather than evidence that the global prior is learned from data. Please state this interpretation explicitly.
  2. [Table 1 and Figure 5] No standard deviations or numbers of seeds are reported in Table 1, and Figure 5's legend lists 'PerFed' while the tables use 'pFedMe'; please fix the label and report variability.
  3. [Section 5.3.1] The ECE reliability analysis is reported for CelebA only; adding the Dogcat calibration result would strengthen the comparison.
  4. [Appendix C] The notation x_n is used ambiguously: Eq. (13) calls x_n predictive points, while Eq. (14) uses the same notation in likelihood terms for training data; please define the index sets explicitly.
  5. [Section 5.1.1, Dogcat] The regression labels are constructed by adding Gaussian noise to the classification labels, so the two tasks are nearly deterministic functions of each other; this is an artificial correlation that may overstate the benefit of MOGP transfer on this benchmark.
  6. [Appendix D] The ELBO optimization in Eq. (19) uses AdamW for kernel hyperparameters and mixing weights, while Eq. (20) gives a closed form for the noise variances; the paper does not report the learning rate or optimizer hyperparameters, which limits reproducibility.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the pFed-Mul derivation is self-contained, and the cited inducing-point approximation is an explicit modeling choice rather than a fitted target renamed as a prediction.

full rationale

Walking the derivation chain: the local variational updates in Eq. (3) follow from the MOGP likelihood (Eq. 1) via the Pólya-Gamma augmentation identity of Polson et al. (external and independently checkable), and the predictive distribution in Eq. (4) is the standard GP conditional, not a re-statement of a fitted constant. The server update in Eq. (5) maximizes an averaged ELBO over kernel hyperparameters, mixing weights, and noise variances on the clients' training data; this is empirical Bayes, and the reported MSE/ACC/ECE/OOD numbers are computed on held-out or masked inputs, so no prediction is statistically forced by a fitted parameter. The synthetic experiment initializes hyperparameters to the ground truth, but it is a recovery sanity check, not a prediction claim. The one same-author citation, [68], is used in Appendix C to justify the deterministic inducing-point projection (Eq. 13); this DTC-style projection can understate posterior variance and is a legitimate correctness/calibration risk, but it is openly stated as a simplification, is externally falsifiable, and does not make any experimental result equal to its input by construction. No load-bearing argument reduces to a self-citation or to a redefinition of the target quantity.

Assumptions & free parameters 6 free parameters · 7 assumptions · 0 invented entities

The method's support rests on several fitted quantities (W, kernel hyperparameters, deep backbone, noise variances, B, inducing points). These are estimated from the same training data used for evaluation, which is empirical Bayes; the prior is therefore not fixed a priori. The main ad hoc modeling choices are the mean-field factorization and the deterministic projection in Eq (13), which are not derived from data and could distort uncertainty calibration. These assumptions are stated in the paper, so the ledger is transparent.

free parameters (6)
  • Mixing weights W = Tuned on server via ELBO (pFed-Mul-N); in real data W is tuned with fixed other hyperparameters (Sec 5.1.3)
    Defines cross-task correlations in the MOGP prior; fitted to training data rather than fixed.
  • Kernel hyperparameters phi_0, phi_1 per basis kernel = Optimized by Eq (19); synthetic initialized to ground truth
    Amplitude and lengthscale of RBF base kernels; fitted on training data through averaged ELBO.
  • Deep kernel backbone theta (ResNet-18) = Optimized on server for pFed-Mul-N
    Neural network parameters in the deep kernel; fitted by maximizing averaged ELBO.
  • Regression noise variances sigma_i^2 = Closed-form estimate via Eq (20)
    Noise variance in regression likelihood; estimated from the training residuals.
  • Number of basis functions B = B=2 in all experiments
    Rank of the coregionalization matrix; chosen by hand, not validated by model selection.
  • Inducing point locations and count M = Uniformly sampled from local data; M not reported
    Sparse GP approximation quality depends on M and on the sampling scheme; neither is optimized nor fully specified.
assumptions (7)
  • standard math Pólya-Gamma augmentation identity (Theorem 1 of Polson et al. 2013)
    Used in Appendix A to convert the logistic classification likelihood into a conditionally conjugate Gaussian form.
  • domain assumption Mean-field factorization q(omega,f) = q1(omega) q2(f)
    Assumed in Section 4.1.2; ignores posterior dependence between Pólya-Gamma variables and latent functions.
  • domain assumption Linear model of coregionalization with B shared basis functions
    Section 3.2 and 4.1.1; each output is a linear combination of the same B independent Gaussian processes, a low-rank assumption on cross-task covariance.
  • ad hoc to paper Deterministic conditional approximation in Eq (13)
    Appendix C replaces the conditional distribution of latent values given inducing outputs with its conditional mean, discarding conditional variance.
  • domain assumption Inducing inputs uniformly sampled from local data
    Section 4.3; assumes random uniform subsampling gives adequate inducing points, and these points are not uploaded to the server.
  • domain assumption Identical dataset size across clients
    Section 4 states 'For convenience, we assume an identical dataset size across all clients'; real deployments may violate this.
  • domain assumption Averaged ELBO maximization yields a useful global prior
    Section 4.2; assumes that optimizing the average ELBO over fixed local variational posteriors improves the global MOGP prior for all clients.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Task Diversity in Bayesian Federated Learning: Simultaneous Processing of Classification and Regression." pith.science (2026). https://pith.science/paper/AIRVWDUO

@misc{pith2026241210897,
  author       = {Pith},
  title        = {Pith review of: Task Diversity in Bayesian Federated Learning: Simultaneous Processing of Classification and Regression},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/AIRVWDUO}},
  note         = {Machine review of arXiv:2412.10897}
}
read the original abstract

This work addresses a key limitation in current federated learning approaches, which predominantly focus on homogeneous tasks, neglecting the task diversity on local devices. We propose a principled integration of multi-task learning using multi-output Gaussian processes (MOGP) at the local level and federated learning at the global level. MOGP handles correlated classification and regression tasks, offering a Bayesian non-parametric approach that naturally quantifies uncertainty. The central server aggregates the posteriors from local devices, updating a global MOGP prior redistributed for training local models until convergence. Challenges in performing posterior inference on local devices are addressed through the P\'{o}lya-Gamma augmentation technique and mean-field variational inference, enhancing computational efficiency and convergence rate. Experimental results on both synthetic and real data demonstrate superior predictive performance, OOD detection, uncertainty calibration and convergence rate, highlighting the method's potential in diverse applications. Our code is publicly available at https://github.com/JunliangLv/task_diversity_BFL.

Figures

Figures reproduced from arXiv: 2412.10897 by the authors.

Figure 1
Figure 1. The overview of our model pFed-Mul. Left: System diagram. The central server aggregates the posteriors from local [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. The estimated posterior of latent functions from [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Reliability diagrams for all methods. We plot the perfect calibration as blue diagonals, and practical result as orange [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (1 more)
Figure 5
Figure 5. Figure 5: Convergence rate of all models on both datasets. [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

74 extracted references · 56 canonical work pages

  1. [1]

    Idan Achituve, Aviv Shamsian, Aviv Navon, Gal Chechik, and Ethan Fetaya. 2021. Personalized federated learning with gaussian processes. Advances in Neural Information Processing Systems 34 (2021), 8392– 8406

    work page 2021

  2. [2]

    Naman Agarwal, Ananda Theertha Suresh, Felix Xinnan X Yu, San- jiv Kumar, and Brendan McMahan. 2018. cpSGD: Communication- efficient and differentially-private distributed SGD. Advances in Neural Information Processing Systems 31 (2018)

    work page 2018

  3. [3]

    Álvarez, Lorenzo Rosasco, and Neil D

    Mauricio A. Álvarez, Lorenzo Rosasco, and Neil D. Lawrence. 2012. Kernels for Vector-Valued Functions: A Review.Found. Trends Mach. Learn. 4, 3 (2012), 195–266

    work page 2012

  4. [4]

    Shun-Ichi Amari. 1998. Natural gradient works efficiently in learning. Neural computation 10, 2 (1998), 251–276

    work page 1998

  5. [5]

    Manoj Ghuhan Arivazhagan, Vinay Aggarwal, Aaditya Kumar Singh, and Sunav Choudhary. 2019. Federated learning with personalization layers. arXiv preprint arXiv:1912.00818 (2019)

    arXiv 2019

  6. [6]

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

    work page 2017

  7. [7]

    Longbing Cao, Hui Chen, Xuhui Fan, Joao Gama, Yew-Soon Ong, and Vipin Kumar. 2023. Bayesian Federated Learning: A Survey. arXiv preprint arXiv:2304.13267 (2023)

    arXiv 2023

  8. [8]

    Rich Caruana. 1997. Multitask learning. Machine learning 28 (1997), 41–75

    1997

Show all 74 references
  1. [9]

    Mingzhe Chen, Nir Shlezinger, H Vincent Poor, Yonina C Eldar, and Shuguang Cui. 2021. Communication-efficient federated learn- ing. Proceedings of the National Academy of Sciences 118, 17 (2021), e2024789118

    2021

  2. [10]

    Ronan Collobert and Jason Weston. 2008. A unified architecture for natural language processing: Deep neural networks with multitask learning. In Proceedings of the 25th international conference on Machine learning. 160–167

    2008

  3. [11]

    Luca Corinzia, Ami Beuret, and Joachim M Buhmann. 2019. Variational federated multi-task learning. arXiv preprint arXiv:1906.06268 (2019)

    2019 arXiv

  4. [12]

    Zhongxiang Dai, Bryan Kian Hsiang Low, and Patrick Jaillet. 2020. Federated Bayesian optimization via Thompson sampling. Advances in Neural Information Processing Systems 33 (2020), 9687–9699

    2020

  5. [13]

    Canh T Dinh, Tung T Vu, Nguyen H Tran, Minh N Dao, and Hongyu Zhang. 2021. Fedu: A unified framework for federated multi-task learning with laplacian regularization. arXiv preprint arXiv:2102.07148 400 (2021)

    2021 arXiv

  6. [14]

    Daxiang Dong, Hua Wu, Wei He, Dianhai Yu, and Haifeng Wang. 2015. Multi-task learning for multiple language translation. In Proceedings of the 53rd Annual Meeting of the Association for Computational Lin- guistics and the 7th International Joint Conference on Natural Language ...

    2015

  7. [15]

    Alireza Fallah, Aryan Mokhtari, and Asuman Ozdaglar. 2020. On the convergence theory of gradient-based model-agnostic meta-learning algorithms. In International Conference on Artificial Intelligence and Statistics. PMLR, 1082–1092

    2020

  8. [16]

    Alireza Fallah, Aryan Mokhtari, and Asuman Ozdaglar. 2020. Person- alized federated learning: A meta-learning approach. arXiv preprint arXiv:2002.07948 (2020)

    2020 arXiv

  9. [17]

    Théo Galy-Fajou, Florian Wenzel, Christian Donner, and Manfred Op- per. 2020. Multi-class gaussian process classification made conjugate: Efficient inference via data augmentation. In Uncertainty in Artificial Intelligence. PMLR, 755–765

    2020

  10. [18]

    Min Gao, Jian-Yu Li, Chun-Hua Chen, Yun Li, Jun Zhang, and Zhi-Hui Zhan. 2023. Enhanced multi-task learning and knowledge graph- based recommender system. IEEE Transactions on Knowledge and Data Engineering (2023)

    2023

  11. [19]

    Chuan Guo, Geoff Pleiss, Yu Sun, and Kilian Q Weinberger. 2017. On calibration of modern neural networks. In International Conference on Machine Learning. PMLR, 1321–1330

    2017

  12. [20]

    Farzin Haddadpour and Mehrdad Mahdavi. 2019. On the conver- gence of local descent methods in federated learning. arXiv preprint arXiv:1910.14425 (2019)

    2019 arXiv

  13. [21]

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. 2016. Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition . 770–778. KDD ’25, August 3–7, 2025, Toronto, ON, Canada Junliang Lyu, Yixuan Zhang, Xia...

    2016

  14. [22]

    James Hensman, Alexander Matthews, and Zoubin Ghahramani. 2015. Scalable variational Gaussian process classification. In Artificial Intel- ligence and Statistics. PMLR, 351–360

    2015

  15. [23]

    Matthew D Hoffman, David M Blei, Chong Wang, and John Paisley

  16. [24]

    Yutao Huang, Lingyang Chu, Zirui Zhou, Lanjun Wang, Jiangchuan Liu, Jian Pei, and Yong Zhang. 2021. Personalized cross-silo federated learning on non-iid data. In Proceedings of the AAAI conference on artificial intelligence, Vol. 35. 7865–7873

    2021

  17. [25]

    Salman Jahani, Shiyu Zhou, Dharmaraj Veeramani, and Jeff Schmidt

  18. [26]

    Yihan Jiang, Jakub Konečn`y, Keith Rush, and Sreeram Kannan. 2019. Improving federated learning personalization via model agnostic meta learning. arXiv preprint arXiv:1909.12488 (2019)

    2019 arXiv

  19. [27]

    Andre G Journel and Charles J Huijbregts. 1976. Mining geostatistics. Academic Press

    1976

  20. [28]

    Sai Praneeth Karimireddy, Satyen Kale, Mehryar Mohri, Sashank Reddi, Sebastian Stich, and Ananda Theertha Suresh. 2020. Scaffold: Sto- chastic controlled averaging for federated learning. In International conference on machine learning . PMLR, 5132–5143

    2020

  21. [29]

    Tianjun Ke, Haoqun Cao, Zenan Ling, and Feng Zhou. 2023. Revisiting Logistic-softmax Likelihood in Bayesian Meta-Learning for Few-Shot Classification. arXiv preprint arXiv:2310.10379 (2023)

    2023 arXiv

  22. [30]

    Hui Li, Yanlin Wang, Ziyu Lyu, and Jieming Shi. 2020. Multi-task learning for recommendation over heterogeneous information net- work. IEEE Transactions on Knowledge and Data Engineering 34, 2 (2020), 789–802

    2020

  23. [31]

    Rui Li, Fenglong Ma, Wenjun Jiang, and Jing Gao. 2019. Online feder- ated multitask learning. In 2019 IEEE International Conference on Big Data (Big Data). IEEE, 215–220

    2019

  24. [32]

    Tian Li, Anit Kumar Sahu, Maziar Sanjabi, Manzil Zaheer, Ameet Tal- walkar, and Virginia Smith. 2018. On the convergence of federated op- timization in heterogeneous networks. arXiv preprint arXiv:1812.06127 (2018)

    2018 arXiv

  25. [33]

    Liangxi Liu, Xi Jiang, Feng Zheng, Hong Chen, Guo-Jun Qi, Heng Huang, and Ling Shao. 2023. A bayesian federated learning frame- work with online laplace approximation. IEEE Transactions on Pattern Analysis and Machine Intelligence (2023)

    2023

  26. [34]

    Shikun Liu, Edward Johns, and Andrew J Davison. 2019. End-to-end multi-task learning with attention. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition . 1871–1880

    2019

  27. [35]

    Ziwei Liu, Ping Luo, Xiaogang Wang, and Xiaoou Tang. 2015. Deep Learning Face Attributes in the Wild. In Proceedings of International Conference on Computer Vision (ICCV)

    2015

  28. [36]

    Yong Luo, Dacheng Tao, Bo Geng, Chao Xu, and Stephen J Maybank

  29. [37]

    Ningning Ma, Xiangyu Zhang, Hai-Tao Zheng, and Jian Sun. 2018. ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design. arXiv:1807.11164 [cs.CV]

    2018 arXiv

  30. [38]

    Othmane Marfoq, Giovanni Neglia, Aurélien Bellet, Laetitia Kameni, and Richard Vidal. 2021. Federated multi-task learning under a mixture of distributions. Advances in Neural Information Processing Systems 34 (2021), 15434–15447

    2021

  31. [39]

    Brendan McMahan, Eider Moore, Daniel Ramage, Seth Hampson, and Blaise Aguera y Arcas. 2017. Communication-efficient learning of deep networks from decentralized data. In Artificial intelligence and statistics. PMLR, 1273–1282

    2017

  32. [40]

    Pablo Moreno-Muñoz, Antonio Artés, and Mauricio Alvarez. 2018. Heterogeneous multi-output Gaussian process prediction. Advances in neural information processing systems 31 (2018)

    2018

  33. [41]

    Yongli Mou, Jiahui Geng, Feng Zhou, Oya Beyan, Chunming Rong, and Stefan Decker. 2023. pFedV: Mitigating Feature Distribution Skewness via Personalized Federated Learning with Variational Distribution Constraints. In Pacific-Asia Conference on Knowledge Discovery and Data Mini...

    2023

  34. [42]

    Radford M Neal. 1993. Probabilistic inference using Markov chain Monte Carlo methods. Department of Computer Science, University of Toronto Toronto, ON, Canada

    1993

  35. [43]

    Nicholas G Polson, James G Scott, and Jesse Windle. 2013. Bayesian inference for logistic models using Pólya-Gamma latent variables. Journal of the American statistical Association 108, 504 (2013), 1339– 1349

    2013

  36. [44]

    Ilija Radosavovic, Raj Prateek Kosaraju, Ross Girshick, Kaiming He, and Piotr Dollár. 2020. Designing Network Design Spaces. arXiv:2003.13678 [cs.CV]

    2020 arXiv

  37. [45]

    Carl Edward Rasmussen. 2003. Gaussian processes in machine learning. In Summer School on Machine Learning . Springer, 63–71

    2003

  38. [46]

    Amirhossein Reisizadeh, Aryan Mokhtari, Hamed Hassani, Ali Jad- babaie, and Ramtin Pedarsani. 2020. Fedpaq: A communication- efficient federated learning method with periodic averaging and quanti- zation. In International Conference on Artificial Intelligence and Statistics. P...

    2020

  39. [47]

    Daniel Rothchild, Ashwinee Panda, Enayat Ullah, Nikita Ivkin, Ion Stoica, Vladimir Braverman, Joseph Gonzalez, and Raman Arora. 2020. Fetchsgd: Communication-efficient federated learning with sketching. In International Conference on Machine Learning . PMLR, 8253–8265

    2020

  40. [48]

    Sebastian Ruder. 2017. An overview of multi-task learning in deep neural networks. arXiv preprint arXiv:1706.05098 (2017)

    2017 arXiv

  41. [49]

    Felix Sattler, Simon Wiedemann, Klaus-Robert Müller, and Wojciech Samek. 2019. Robust and communication-efficient federated learning from non-iid data. IEEE transactions on neural networks and learning systems 31, 9 (2019), 3400–3413

    2019

  42. [50]

    Virginia Smith, Chao-Kai Chiang, Maziar Sanjabi, and Ameet S Tal- walkar. 2017. Federated multi-task learning. Advances in neural infor- mation processing systems 30 (2017)

    2017

  43. [51]

    Jake Snell and Richard Zemel. 2020. Bayesian Few-Shot Classification with One-vs-Each P\’olya-Gamma Augmented Gaussian Processes. arXiv preprint arXiv:2007.10417 (2020)

    2020 arXiv

  44. [52]

    Sebastian U Stich. 2018. Local SGD Converges Fast and Communicates Little. In International Conference on Learning Representations

    2018

  45. [53]

    Canh T Dinh, Nguyen Tran, and Josh Nguyen. 2020. Personalized feder- ated learning with moreau envelopes. Advances in Neural Information Processing Systems 33 (2020), 21394–21405

    2020

  46. [54]

    Alysa Ziying Tan, Han Yu, Lizhen Cui, and Qiang Yang. 2022. Towards personalized federated learning. IEEE Transactions on Neural Networks and Learning Systems (2022)

    2022

  47. [55]

    Mingxing Tan and Quoc Le. 2019. Efficientnet: Rethinking model scaling for convolutional neural networks. In International conference on machine learning. PMLR, 6105–6114

    2019

  48. [56]

    Michalis Titsias. 2009. Variational learning of inducing variables in sparse Gaussian processes. In Artificial Intelligence and Statistics. 567– 574

    2009

  49. [57]

    Aleksei Triastcyn and Boi Faltings. 2019. Federated learning with bayesian differential privacy. In 2019 IEEE International Conference on Big Data (Big Data) . IEEE, 2587–2596

    2019

  50. [58]

    Stacey Truex, Nathalie Baracaldo, Ali Anwar, Thomas Steinke, Heiko Ludwig, Rui Zhang, and Yi Zhou. 2019. A hybrid approach to privacy- preserving federated learning. InProceedings of the 12th ACM workshop on artificial intelligence and security . 1–11

    2019

  51. [59]

    Kang Wei, Jun Li, Ming Ding, Chuan Ma, Howard H Yang, Farhad Farokhi, Shi Jin, Tony QS Quek, and H Vincent Poor. 2020. Federated Task Diversity in Bayesian Federated Learning: Simultaneous Processing of Classification and Regression KDD ’25, August 3–7, 2025, Toronto, ON, Cana...

    2020

  52. [60]

    Florian Wenzel, Théo Galy-Fajou, Christan Donner, Marius Kloft, and Manfred Opper. 2019. Efficient Gaussian process classification using Pólya-Gamma data augmentation. In Proceedings of the AAAI Confer- ence on Artificial Intelligence, Vol. 33. 5417–5424

    2019

  53. [61]

    Andrew Gordon Wilson, Zhiting Hu, Ruslan Salakhutdinov, and Eric P Xing. 2016. Deep kernel learning. In Artificial intelligence and statistics. PMLR, 370–378

    2016

  54. [62]

    Jian Xu, Xinyi Tong, and Shao-Lun Huang. 2023. Personalized fed- erated learning with feature alignment and classifier collaboration. arXiv preprint arXiv:2306.11867 (2023)

    2023 arXiv

  55. [63]

    Qiang Yang, Yang Liu, Tianjian Chen, and Yongxin Tong. 2019. Feder- ated machine learning: Concept and applications. ACM Transactions on Intelligent Systems and Technology (TIST) 10, 2 (2019), 1–19

    2019

  56. [64]

    Feng Yin, Zhidi Lin, Qinglei Kong, Yue Xu, Deshi Li, Sergios Theodor- idis, and Shuguang Robert Cui. 2020. FedLoc: Federated learning framework for data-driven cooperative localization and location data processing. IEEE Open Journal of Signal Processing 1 (2020), 187–215

    2020

  57. [65]

    Haolin Yu, Kaiyang Guo, Mahdi Karami, Xi Chen, Guojun Zhang, and Pascal Poupart. 2022. Federated Bayesian Neural Regression: A Scalable Global Federated Gaussian Process. arXiv preprint arXiv:2206.06357 (2022)

    2022 arXiv

  58. [66]

    Chen Zhang, Yu Xie, Hang Bai, Bin Yu, Weihong Li, and Yuan Gao

  59. [67]

    Xu Zhang, Yinchuan Li, Wenpeng Li, Kaiyang Guo, and Yunfeng Shao

  60. [68]

    Feng Zhou, Quyu Kong, Zhijie Deng, Fengxiang He, Peng Cui, and Jun Zhu. 2023. Heterogeneous multi-task Gaussian Cox processes. Machine Learning (2023), 1–30

    2023

  61. [69]

    Junyi Zhu, Xingchen Ma, and Matthew B Blaschko. 2023. Confidence- aware personalized federated learning via variational expectation max- imization. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 24542–24551. KDD ’25, August 3–7, 2025, Toro...

    2023

  62. [70]

    Knowledge-Based Systems 216 (2021), 106775

    A survey on federated learning. Knowledge-Based Systems 216 (2021), 106775

    2021

  63. [2012]

    IEEE Transactions on Image Processing 22, 2 (2012), 523–536

    Manifold regularized multitask learning for semi-supervised multilabel image classification. IEEE Transactions on Image Processing 22, 2 (2012), 523–536

    2012

  64. [2013]

    Journal of Machine Learning Research (2013)

    Stochastic variational inference. Journal of Machine Learning Research (2013)

    2013

  65. [2021]

    IEEE Transactions on Reliability (2021)

    Multioutput Gaussian Process Modulated Poisson Processes for Event Prediction. IEEE Transactions on Reliability (2021)

    2021

  66. [2022]

    In International Conference on Machine Learning

    Personalized federated learning via variational bayesian infer- ence. In International Conference on Machine Learning . PMLR, 26293– 26310

Pith tools

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