Pith. sign in

REVIEW 3 major objections 5 minor 64 references

Lazy But Effective: Collaborative Personalized Federated Learning with Heterogeneous Data

T0 review · 3 major / 5 minor · reviewed 2026-08-16 · deepseek-v4-flash

Pith's one-line read A lazy influence score clusters federated clients with oracle-level accuracy.

desk verdict Plausible clustering method for personalized FL whose Oracle-matching claim rests on an unvalidated warm-up operating point. read the letter →

arxiv 2505.02540 v1 pith:25F3SWJP submitted 2025-05-05 cs.LG cs.AI

classification cs.LGcs.AI
keywords personalizedfederatedlearningclientclusteringlazyinfluenceapproximationnon-IIDdatalabelheterogeneityfeaturepeer-to-peernext-wordprediction
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 tries to establish that a cheap, one-time estimate of how much each client's data helps every other client's validation loss is enough to partition federated-learning clients into groups that should train together, without knowing how many groups exist. This matters because a single global model trained over non-IID client data performs poorly for individual users, while per-client models lose the benefit of collaboration. pFedLIA computes a 'lazy influence' score for every pair of clients, clusters with OPTICS in the centralized case or with k-means in the peer-to-peer case, and then runs any federated aggregator inside each cluster. The paper reports that this matches the accuracy of an oracle cluster assignment and beats the best baseline by up to 17% on CIFAR100.

What carries the argument

The Lazy Influence Approximation (LIA) is the mechanism: a pairwise score between clients defined as the reduction in client $i$'s validation loss when the shared warm-up model is moved by a few local training steps on client $j$'s data. It is a deliberately inaccurate but cheap proxy for exact influence, meant to preserve only the direction and ordering of true influence. Clustering on this matrix, either centrally with OPTICS, which needs no target cluster count in advance, or per-client with a two-cluster k-means split, is what converts the scores into personalized training groups.

What would settle it

Take two clusters with disjoint label sets, compute the LIA matrix with the paper's recommended short run, and check whether every within-cluster score exceeds every between-cluster score. A single inversion of that ordering, visible as a positive $I_{LIA}$ between clients from different clusters, would break the clustering step and prevent the method from matching the oracle.

Watch

Extended reading notes

Core claim

The central claim is that the sign and ordering of $I_{LIA}(i,j) = \sum_{z \in Z^{\mathrm{val}}_i} L(z,\theta_0) - L(z,\theta_j)$ carry enough distributional signal to recover the performance lost to data heterogeneity. Here $\theta_j$ is obtained by a small number of local training epochs on client $j$'s batch starting from a shared warm-up model $\theta_0$, and $Z^{\mathrm{val}}_i$ is client $i$'s validation set. The paper argues that exact influence is too costly and that first-order influence approximations require Hessian operations that are impractical for modern models, whereas this lazy score needs only a few local updates and one validation pass. It then shows that clustering once on this score, before any cluster-specific training, matches a hypothetical Oracle clustering and outperforms existing clustering and personalization baselines across pathological and noisy non-IID splits of image benchmarks and a Nordic-language next-word prediction task.

Load-bearing premise

The load-bearing premise is that the small loss difference observed after a few local training steps from a shared starting model ranks other clients by how much their data genuinely helps the client's task; if that ordering is an optimization artifact, the clusters will be wrong.

Editorial extensions

If this is right

  • Within a small number of communication rounds, pFedLIA can switch from a single global model to cluster-specific models, so personalization does not require iterative re-clustering.
  • Because the clustering step needs no prior knowledge of the number of clusters, the method applies to federations whose client distribution is unknown in structure.
  • The same lazy score can be computed in a peer-to-peer setting, so each client can choose its own aggregation partners rather than relying on a central server.
  • The reported 40x to 500x speedup over exact influence makes the method practical for large models while preserving most of the personalization benefit.
  • On the pathological and noisy non-IID benchmarks, the method matches Oracle accuracy and outperforms the strongest baseline by up to 17%, demonstrating that a one-time clustering step can recover much of the performance lost to heterogeneity.

Reading between the lines

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

  • An extension the authors leave implicit: if the LIA ordering remains stable under distribution shift, the same one-time clustering could be re-run periodically, turning pFedLIA into a lightweight adaptive personalization scheme.
  • Because the score is computed from a small validation set, clients with very few labeled validation samples are a risk case; a testable extension is to measure how cluster quality degrades as validation set size shrinks.
  • The paper demonstrates the method with FedAvg as the within-cluster aggregator; nothing in the argument prevents combining the same clusters with robust aggregation methods, which could additionally filter malicious or outlier clients.
  • Another inference: the lazy score only needs the sign and ordering of influence, so the method's failure mode is not miscalibrated loss values but inconsistent ranking across initializations; a stress test would randomize the warm-up model and check cluster stability.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper proposes pFedLIA, a personalized federated learning method that clusters clients before model aggregation using a Lazy Influence Approximation (LIA). Each client j performs a small number of local epochs from a common reference model θ0 on a batch of its own data, sends the partially updated model to client i, and i scores j by the reduction in i's validation loss (Eq. 1). Clustering is performed once, either centrally with OPTICS or peer-to-peer with k-means, without requiring the number of clusters in advance; subsequent training uses any federated aggregator within each cluster. The method is evaluated on FashionMNIST, CIFAR10, CIFAR100 under pathological and noisy non-IID splits, and on a Scandinavian next-word prediction task with GPT-2. The headline claims are that pFedLIA matches the performance of an Oracle clustering and outperforms the best baseline by up to 17%.

Significance. If the central claim holds, the paper offers a simple and communication-efficient way to obtain personalized models in heterogeneous federated learning, with the practical advantage of a one-time clustering step and no need to know the number of clusters. The evaluation spans multiple datasets and modalities, includes a real-world language task, and compares against several established baselines with mean and standard deviation reported over four runs. The main weakness is that the core quantity, the Lazy Influence score, is not validated as a faithful proxy for the exact influence defined in Section III-B, and the headline Oracle-matching result is demonstrated only in the clean pathological setting. These issues are load-bearing for the stated contributions.

major comments (3)
  1. [§III-D, Eq. (1), and §IV-D] The quantity computed in the evaluation is not the influence defined in Section III-B. In the experiments, θ0 is the FedAvg model after 20 communication rounds on all clients' data, so every client's data already contributes to θ0. Thus I_LIA(i,j) measures the effect of a few extra local steps from a model that already contains j's contribution, not the marginal effect of adding j's data to a model trained without it. A first-order expansion shows I_LIA(i,j) is approximately a gradient-alignment score (an inner product of the validation gradient and the training-batch gradient). The paper provides no ablation over the warm-up length (round 20) or the number of local epochs k, and no cluster-purity or score-fidelity metric. The 'matches Oracle' result could therefore be an artifact of a specific operating point rather than a robust property of the proposed approximation. This directly affects the central claim and should be addressed with a calibration experiment, an ablation, or a theoretical condition under which the score preserves the sign and ordering of the exact influence.
  2. [Table I and Abstract] The claim that pFedLIA 'matches the performance of the Oracle clustering' is supported only for the pathological non-IID setting, where labels are exclusively assigned to clusters and perfect recovery is expected. The Oracle row is absent for the noisy non-IID setting, so the claim is not established there. In noisy FashionMNIST, pFedLIA is actually worse than FedFomo (85.11% vs. 82.10%) and FedProto (87.39% vs. 82.10%), which conflicts with the unqualified wording of the abstract and conclusion. The large standard deviations, e.g., 57.09% ± 13.32% on noisy CIFAR100, also make the 17% improvement fragile. Please qualify the Oracle-matching claim to the setting where an Oracle is defined and add the missing comparison or a clear statement of scope.
  3. [§III-D] The core assumption of the method is that a small number of noisy local updates from θ0, combined with a small validation set, preserves the sign and ordering of the true influence of one client's data on another. The paper states that 'we only need an accurate enough estimate of its direction' but provides no theoretical guarantee and no empirical validation of this property. Figure 3a compares LIA with exact influence only in computation time, not in the fidelity of the resulting scores. A calibration experiment on a small problem where exact influence is tractable, or a study of cluster purity under varying k and warm-up length, is necessary to establish that the clustering signal is actually the influence signal claimed.
minor comments (5)
  1. [§IV-B] The noisy non-IID description says 'we opted to only add 1' random label per client but then states that CIFAR100 assigns five random labels instead; please clarify the protocol used for each dataset.
  2. [§III-D vs. §IV-D] The method description says the warm-up model M0 'can be trained, e.g., by the center C using a small set of warm-up data,' while the evaluation trains it via FedAvg on all clients' data for 20 rounds. Please specify which protocol the reported results use and discuss whether the warm-up model must exclude target data for the influence interpretation to hold.
  3. [Figure 3a] The exact-influence comparison uses three convergence thresholds with no explanation of how they were chosen; adding the threshold values and error bars would make the speedup claim more reproducible.
  4. [§IV-D] Reporting four runs without statistical tests is insufficient given the large standard deviations in Table I; consider reporting confidence intervals or performing a significance test for the headline improvements.
  5. [Abstract] The phrase 'an improvement of 17% on CIFAR100' refers to the noisy setting, while the pathological CIFAR100 improvement is 16.93% over FeSEM; please identify the exact comparison in the text to avoid ambiguity.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: pFedLIA's clustering score is defined directly in Eq. (1) and validated on external benchmarks, so the central claim does not reduce to its inputs.

full rationale

The central claim that pFedLIA matches an Oracle clustering and improves on baselines is an empirical result obtained on external datasets (FashionMNIST, CIFAR10, CIFAR100, OSCAR) against external baselines (FedAvg, FedFomo, FeSEM, CFL, FedProto, Local-only). Equation (1) defines the Lazy Influence score directly as a difference of validation losses; no parameter is fitted to the reported accuracies or to the cluster labels, and the Oracle is an upper-bound baseline rather than a term in the score. The Lazy Influence approximation is taken from the authors' prior work [43], which is a self-citation, but it is not the sole justification for the empirical claim: the ability of the score to cluster clients is tested here on held-out benchmarks, so the citation is not load-bearing in a way that makes the derivation circular. The paper's methodological weaknesses (e.g., the warm-up model at round 20 already contains all clients' data, making Eq. (1) closer to a gradient-similarity heuristic than to exact influence, and the absence of ablations over the warm-up length or k) are correctness and robustness concerns, not circularity: they do not show that any 'prediction' is equivalent to an input by construction. No uniqueness theorem, ansatz smuggled through citation, or renaming of a known result as organization was found. Therefore the derivation chain is self-contained with respect to circularity.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

No invented physical or conceptual entities are introduced; the method reuses influence functions, clustering algorithms, and FedAvg. The main unverified premise is that a cheap partial-training influence score is a sufficient statistic for clustering. The free parameters are standard hyperparameters, but several are unreported, which limits reproducibility.

free parameters (4)
  • LIA local epochs k = 20
    Section III-D: each client performs k local epochs to produce the partial model theta_j. This is a hand-chosen hyperparameter that affects the quality of the influence estimate.
  • Warm-up communication rounds = 20
    Section IV-D: pFedLIA is applied after the 20th communication round of FedAvg. This choice follows CFL but is a hyperparameter of the evaluation and may affect clustering quality.
  • OPTICS hyperparameters = unspecified
    Section III-F: centralized clustering uses OPTICS, but min_samples, xi, and eps are not reported in this version. Results depend on these unstated settings.
  • LIA batch size per client = unspecified
    Section III-D: each client uses a batch of its training data to compute the partial model; the batch size is not reported, despite being critical to the influence estimate.
assumptions (5)
  • domain assumption Clients can be naturally partitioned into clusters with similar data distributions.
    Stated in Section I; the entire approach depends on clusterable structure in the client population.
  • ad hoc to paper The Lazy Influence Approximation in Eq. (1), using a few local epochs, preserves the sign and ordering of the true influence of one client's data on another.
    No proof is given that a partially trained model provides a reliable similarity signal; this is the load-bearing assumption behind the clustering.
  • domain assumption Each client's validation set is representative of its data distribution.
    Used in Eq. (1) to compute I_LIA(i,j); if the validation split is small or skewed, the influence scores are noisy.
  • standard math The empirical risk is differentiable.
    Assumed in Section III-A and required for gradient-based local updates.
  • domain assumption Pretrained model initialization for ResNet18 and GPT-2 does not distort the influence signal.
    Section IV-D states pretrained weights are used; the effect of initialization on clustering quality is not tested.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Lazy But Effective: Collaborative Personalized Federated Learning with Heterogeneous Data." pith.science (2026). https://pith.science/paper/25F3SWJP

@misc{pith2026250502540,
  author       = {Pith},
  title        = {Pith review of: Lazy But Effective: Collaborative Personalized Federated Learning with Heterogeneous Data},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/25F3SWJP}},
  note         = {Machine review of arXiv:2505.02540}
}
read the original abstract

In Federated Learning, heterogeneity in client data distributions often means that a single global model does not have the best performance for individual clients. Consider for example training a next-word prediction model for keyboards: user-specific language patterns due to demographics (dialect, age, etc.), language proficiency, and writing style result in a highly non-IID dataset across clients. Other examples are medical images taken with different machines, or driving data from different vehicle types. To address this, we propose a simple yet effective personalized federated learning framework (pFedLIA) that utilizes a computationally efficient influence approximation, called `Lazy Influence', to cluster clients in a distributed manner before model aggregation. Within each cluster, data owners collaborate to jointly train a model that captures the specific data patterns of the clients. Our method has been shown to successfully recover the global model's performance drop due to the non-IID-ness in various synthetic and real-world settings, specifically a next-word prediction task on the Nordic languages as well as several benchmark tasks. It matches the performance of a hypothetical Oracle clustering, and significantly improves on existing baselines, e.g., an improvement of 17% on CIFAR100.

Figures

Figures reproduced from arXiv: 2505.02540 by the authors.

Figure 1
Figure 1. Validation loss over time, when training a GPT-2 model [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Visual illustration of the proposed pFedLIA, from [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. (3a) Computation time comparison between the pro￾posed Lazy Influence Approximation (LIA), versus the exact influence (for three different convergence thresholds of vary￾ing precision). LIA achieves between 40–500 times speedup. (3b) pFedLIA decentralized clustering illustration. Consider a classification task (classes [0-7]). We have 32 clients, each assigned 2 out of the 8 classes. This figure depicts the Lazy Inf… view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Training comparison of our proposed method and all baselines on CIFAR100. Note that we train the ‘warm-up’ model [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

64 extracted references · 34 canonical work pages

  1. [1]

    Optics: Ordering points to identify the clustering structure

    Mihael Ankerst, Markus M Breunig, Hans-Peter Kriegel, and J ¨org Sander. Optics: Ordering points to identify the clustering structure. ACM Sigmod record, 28(2):49–60, 1999

  2. [2]

    Federated learning with personalization layers

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

  3. [3]

    Personalized and private peer-to-peer machine learning

    Aur ´elien Bellet, Rachid Guerraoui, Mahsa Taziki, and Marc Tommasi. Personalized and private peer-to-peer machine learning. In International conference on artificial intelligence and statistics . PMLR, 2018

  4. [4]

    Federated learning with hierarchical clustering of local updates to improve training on non- iid data

    Christopher Briggs, Zhong Fan, and Peter Andras. Federated learning with hierarchical clustering of local updates to improve training on non- iid data. arXiv preprint arXiv:2004.11791 , 2020

  5. [5]

    Leaf: A benchmark for federated settings

    Sebastian Caldas, Sai Meher Karthik Duddu, Peter Wu, Tian Li, Jakub Kone ˇcn`y, H Brendan McMahan, Virginia Smith, and Ameet Talwalkar. Leaf: A benchmark for federated settings. arXiv preprint arXiv:1812.01097, 2018

  6. [6]

    On bridging generic and per- sonalized federated learning for image classification

    Hong-You Chen and Wei-Lun Chao. On bridging generic and per- sonalized federated learning for image classification. arXiv preprint arXiv:2107.00778, 2021

  7. [7]

    Federated learning of out-of-vocabulary words

    Mingqing Chen, Rajiv Mathews, Tom Ouyang, and Franc ¸oise Beau- fays. Federated learning of out-of-vocabulary words. arXiv preprint arXiv:1903.10635, 2019

  8. [8]

    Metafed: Federated learning among federations with cyclic knowledge distillation for personalized healthcare

    Yiqiang Chen, Wang Lu, Xin Qin, Jindong Wang, and Xing Xie. Metafed: Federated learning among federations with cyclic knowledge distillation for personalized healthcare. IEEE Transactions on Neural Networks and Learning Systems , 2023

Show all 64 references
  1. [9]

    On robustness properties of convex risk minimization methods for pattern recognition

    Andreas Christmann and Ingo Steinwart. On robustness properties of convex risk minimization methods for pattern recognition. JMLR, 2004

  2. [10]

    Characterizations of an em- pirical influence function for detecting influential cases in regression

    R Dennis Cook and Sanford Weisberg. Characterizations of an em- pirical influence function for detecting influential cases in regression. Technometrics, 1980

  3. [11]

    Residuals and influence in regression

    R Dennis Cook and Sanford Weisberg. Residuals and influence in regression. New York: Chapman and Hall, 1982

  4. [12]

    H˚aler spr ˚aket together the Nordic countries? : a research report on young people’s f ¨ost˚aelse of Danish, Swedish and Norwegian

    Lars-Olof Delsing and Katarina Lundin. H˚aler spr ˚aket together the Nordic countries? : a research report on young people’s f ¨ost˚aelse of Danish, Swedish and Norwegian . Theme North. Nordic Ministers ˚adet,

  5. [13]

    Adaptive personalized federated learning

    Yuyang Deng, Mohammad Mahdi Kamani, and Mehrdad Mah- davi. Adaptive personalized federated learning. arXiv preprint arXiv:2003.13461, 2020

  6. [14]

    Flexible clustered federated learning for client-level data distribution shift

    Moming Duan, Duo Liu, Xinyuan Ji, Yu Wu, Liang Liang, Xianzhang Chen, Yujuan Tan, and Ao Ren. Flexible clustered federated learning for client-level data distribution shift. IEEE Transactions on Parallel and Distributed Systems , 33(11):2661–2674, 2021

  7. [15]

    An efficient framework for clustered federated learning

    Avishek Ghosh, Jichan Chung, Dong Yin, and Kannan Ramchandran. An efficient framework for clustered federated learning. arXiv preprint arXiv:2006.04088, 2020

  8. [16]

    Adaptive aggregation weight assignment for federated learning: A deep reinforcement learning approach

    Enwei Guo, Xiumin Wang, and Weiwei Wu. Adaptive aggregation weight assignment for federated learning: A deep reinforcement learning approach. In Proceedings of the 21st International Conference on Autonomous Agents and Multiagent Systems , pages 1610–1612, 2022

  9. [17]

    Federated learning of a mixture of global and local models

    Filip Hanzely and Peter Richt ´arik. Federated learning of a mixture of global and local models. arXiv preprint arXiv:2002.05516 , 2020

  10. [18]

    Federated learning for mobile keyboard prediction

    Andrew Hard, Kanishka Rao, Rajiv Mathews, Swaroop Ramaswamy, Franc ¸oise Beaufays, Sean Augenstein, Hubert Eichner, Chlo ´e Kiddon, and Daniel Ramage. Federated learning for mobile keyboard prediction. arXiv preprint arXiv:1811.03604 , 2018

  11. [19]

    Fedml: A research library and benchmark for federated machine learning

    Chaoyang He, Songze Li, Jinhyun So, Xiao Zeng, Mi Zhang, Hongyi Wang, Xiaoyang Wang, Praneeth Vepakomma, Abhishek Singh, Hang Qiu, et al. Fedml: A research library and benchmark for federated machine learning. arXiv preprint arXiv:2007.13518 , 2020

  12. [20]

    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 , 2016

  13. [21]

    The non-iid data quagmire of decentralized machine learning

    Kevin Hsieh, Amar Phanishayee, Onur Mutlu, and Phillip B Gibbons. The non-iid data quagmire of decentralized machine learning. arXiv preprint arXiv:1910.00189, 2019

  14. [22]

    Measuring the ef- fects of non-identical data distribution for federated visual classification

    Tzu-Ming Harry Hsu, Hang Qi, and Matthew Brown. Measuring the ef- fects of non-identical data distribution for federated visual classification. arXiv preprint arXiv:1909.06335 , 2019

  15. [23]

    Fedrep: Towards horizontal federated load forecasting for retail energy providers

    Muhammad Akbar Husnoo, Adnan Anwar, Nasser Hosseinzadeh, Shama Naz Islam, Abdun Naser Mahmood, and Robin Doss. Fedrep: Towards horizontal federated load forecasting for retail energy providers. In 2022 IEEE PES 14th Asia-Pacific Power and Energy Engineering Conference (APPEEC)...

  16. [24]

    Advances and open problems in federated learning

    Peter Kairouz, H Brendan McMahan, Brendan Avent, Aur ´elien Bellet, Mehdi Bennis, Arjun Nitin Bhagoji, Kallista Bonawitz, Zachary Charles, Graham Cormode, Rachel Cummings, et al. Advances and open problems in federated learning. Foundations and Trends® in Machine Learning, 14(...

  17. [25]

    Scaffold: Stochastic controlled averaging for federated learning

    Sai Praneeth Karimireddy, Satyen Kale, Mehryar Mohri, Sashank Reddi, Sebastian Stich, and Ananda Theertha Suresh. Scaffold: Stochastic controlled averaging for federated learning. In International Conference on Machine Learning , pages 5132–5143. PMLR, 2020

  18. [26]

    Dynamic clustering in federated learning

    Yeongwoo Kim, Ezeddin Al Hakim, Johan Haraldson, Henrik Eriksson, Jos´e Mairton B da Silva, and Carlo Fischione. Dynamic clustering in federated learning. In ICC 2021-IEEE International Conference on Communications, pages 1–6. IEEE, 2021

  19. [27]

    Understanding black-box predictions via influence functions

    Pang-Wei Koh and Percy Liang. Understanding black-box predictions via influence functions. In International conference on machine learn- ing, pages 1885–1894. PMLR, 2017

  20. [28]

    Learning multiple layers of features from tiny images

    Alex Krizhevsky, Geoffrey Hinton, et al. Learning multiple layers of features from tiny images. Learning Multiple Layers of Features from Tiny Images, 2009

  21. [29]

    Federated learning with soft clustering

    Chengxi Li, Gang Li, and Pramod K Varshney. Federated learning with soft clustering. IEEE Internet of Things Journal, 9(10):7773–7782, 2021

  22. [30]

    Fedmd: Heterogenous federated learning via model distillation

    Daliang Li and Junpu Wang. Fedmd: Heterogenous federated learning via model distillation. arXiv preprint arXiv:1910.03581 , 2019

  23. [31]

    Federated optimization in heterogeneous networks

    Tian Li, Anit Kumar Sahu, Manzil Zaheer, Maziar Sanjabi, Ameet Talwalkar, and Virginia Smith. Federated optimization in heterogeneous networks. Proceedings of Machine Learning and Systems , 2020

  24. [32]

    Fedbn: Federated learning on non-iid features via local batch normal- ization

    Xiaoxiao Li, Meirui Jiang, Xiaofei Zhang, Michael Kamp, and Qi Dou. Fedbn: Federated learning on non-iid features via local batch normal- ization. arXiv preprint arXiv:2102.07623 , 2021

  25. [33]

    Federated learning in mobile edge networks: A comprehensive survey

    Wei Yang Bryan Lim, Nguyen Cong Luong, Dinh Thai Hoang, Yutao Jiao, Ying-Chang Liang, Qiang Yang, Dusit Niyato, and Chunyan Miao. Federated learning in mobile edge networks: A comprehensive survey. IEEE Communications Surveys & Tutorials , 22(3):2031–2063, 2020

  26. [34]

    Efficient approximation of cross-validation for kernel methods using bouligand influence function

    Yong Liu, Shali Jiang, and Shizhong Liao. Efficient approximation of cross-validation for kernel methods using bouligand influence function. In ICML, 2014

  27. [35]

    Multi-center federated learning: clients clustering for better personalization

    Guodong Long, Ming Xie, Tao Shen, Tianyi Zhou, Xianzhi Wang, and Jing Jiang. Multi-center federated learning: clients clustering for better personalization. World Wide Web, 26(1):481–500, 2023

  28. [36]

    Layer-wised model aggregation for personalized federated learning

    Xiaosong Ma, Jie Zhang, Song Guo, and Wenchao Xu. Layer-wised model aggregation for personalized federated learning. In CVPR, 2022

  29. [37]

    Three approaches for personalization with applications to federated learning

    Yishay Mansour, Mehryar Mohri, Jae Ro, and Ananda Theertha Suresh. Three approaches for personalization with applications to federated learning. arXiv preprint arXiv:2002.10619 , 2020

  30. [38]

    Communication-efficient learning of deep networks from decentralized data

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

  31. [39]

    Fedbabu: Towards en- hanced representation for federated image classification

    Jaehoon Oh, Sangmook Kim, and Se-Young Yun. Fedbabu: Towards en- hanced representation for federated image classification. arXiv preprint arXiv:2106.06042, 2021

  32. [40]

    A monolingual approach to contextualized word embeddings for mid- resource languages

    Pedro Javier Ortiz Su’arez, Laurent Romary, and Benoit Sagot. A monolingual approach to contextualized word embeddings for mid- resource languages. In ACL, 2020

  33. [41]

    Clusterfl: A clustering-based federated learning system for human activity recognition

    Xiaomin Ouyang, Zhiyuan Xie, Jiayu Zhou, Guoliang Xing, and Jianwei Huang. Clusterfl: A clustering-based federated learning system for human activity recognition. ACM Transactions on Sensor Networks , 19(1):1–32, 2022

  34. [42]

    Language models are unsupervised multitask learners

    Alec Radford, Jeff Wu, Rewon Child, David Luan, Dario Amodei, and Ilya Sutskever. Language models are unsupervised multitask learners. 2019

  35. [43]

    Lia: Privacy-preserving data quality evaluation in fed- erated learning using a lazy influence approximation

    Ljubomir Rokvic, Panayiotis Danassis, Sai Praneeth Karimireddy, and Boi Faltings. Lia: Privacy-preserving data quality evaluation in fed- erated learning using a lazy influence approximation. In 2024 IEEE International Conference on Big Data (BigData) , 2024

  36. [44]

    Clustered federated learning: Model-agnostic distributed multitask optimization under privacy constraints

    Felix Sattler, Klaus-Robert M ¨uller, and Wojciech Samek. Clustered federated learning: Model-agnostic distributed multitask optimization under privacy constraints. IEEE Transactions on Neural Networks and Learning Systems, 2020

  37. [45]

    Pefll: Per- sonalized federated learning by learning to learn

    Jonathan A Scott, Hossein Zakerinia, and Christoph Lampert. Pefll: Per- sonalized federated learning by learning to learn. In 12th International Conference on Learning Representations , 2024

  38. [46]

    Per- sonalized federated learning using hypernetworks

    Aviv Shamsian, Aviv Navon, Ethan Fetaya, and Gal Chechik. Per- sonalized federated learning using hypernetworks. In International Conference on Machine Learning , pages 9489–9502. PMLR, 2021

  39. [47]

    Federated multi-task learning

    Virginia Smith, Chao-Kai Chiang, Maziar Sanjabi, and Ameet Talwalkar. Federated multi-task learning. CoRR, abs/1705.10467, 2017

  40. [48]

    Personalized fed- erated learning with moreau envelopes

    Canh T Dinh, Nguyen Tran, and Tuan Dung Nguyen. Personalized fed- erated learning with moreau envelopes. Advances in Neural Information Processing Systems, 33, 2020

  41. [49]

    Towards personalized federated learning

    Alysa Ziying Tan, Han Yu, Lizhen Cui, and Qiang Yang. Towards personalized federated learning. IEEE transactions on neural networks and learning systems , 34(12):9587–9603, 2022

  42. [50]

    FL-bench: A federated learning benchmark for solving image classification tasks

    Jiahao Tan and Xinpeng Wang. FL-bench: A federated learning benchmark for solving image classification tasks

  43. [51]

    pfedsim: Similarity-aware model aggregation towards personalized federated learning

    Jiahao Tan, Yipeng Zhou, Gang Liu, Jessie Hui Wang, and Shui Yu. pfedsim: Similarity-aware model aggregation towards personalized federated learning. arXiv preprint arXiv:2305.15706 , 2023

  44. [52]

    Fedproto: Federated prototype learning across heterogeneous clients

    Yue Tan, Guodong Long, Lu Liu, Tianyi Zhou, Qinghua Lu, Jing Jiang, and Chengqi Zhang. Fedproto: Federated prototype learning across heterogeneous clients. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 36, pages 8432–8440, 2022

  45. [53]

    Fuzzy clustered federated learning under mixed data distributions

    Peng Tang, Lifan Wang, Weidong Qiu, Zheng Huang, and Qiangmin Wang. Fuzzy clustered federated learning under mixed data distributions. In Proceedings of the 23rd International Conference on Autonomous Agents and Multiagent Systems , pages 2501–2503, 2024

  46. [54]

    A survey on distributed machine learning

    Joost Verbraeken, Matthijs Wolting, Jonathan Katzy, Jeroen Kloppen- burg, Tim Verbelen, and Jan S Rellermeyer. A survey on distributed machine learning. Acm computing surveys (csur) , 53(2):1–33, 2020

  47. [55]

    A field guide to federated optimization

    Jianyu Wang, Zachary Charles, Zheng Xu, Gauri Joshi, H Brendan McMahan, Maruan Al-Shedivat, Galen Andrew, Salman Avestimehr, Katharine Daly, Deepesh Data, et al. A field guide to federated optimization. arXiv preprint arXiv:2107.06917 , 2021

  48. [56]

    Fashion-mnist: a novel image dataset for benchmarking machine learning algorithms

    Han Xiao, Kashif Rasul, and Roland V ollgraf. Fashion-mnist: a novel image dataset for benchmarking machine learning algorithms. CoRR, abs/1708.07747, 2017

  49. [57]

    Personalized federated learning with feature alignment and classifier collaboration

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

  50. [58]

    Toward understanding the influence of individual clients in federated learning

    Yihao Xue, Chaoyue Niu, Zhenzhe Zheng, Shaojie Tang, Chengfei Lyu, Fan Wu, and Guihai Chen. Toward understanding the influence of individual clients in federated learning. In Proceedings of the AAAI Conference on Artificial Intelligence , 2021

  51. [59]

    Federated machine learning: Concept and applications

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

  52. [60]

    Fedap: Adaptive personaliza- tion in federated learning for non-iid data

    Yousef Yeganeh, Azade Farshad, Johann Boschmann, Richard Gaus, Maximilian Frantzen, and Nassir Navab. Fedap: Adaptive personaliza- tion in federated learning for non-iid data. In International Workshop on Distributed, Collaborative, and Federated Learning , 2022

  53. [61]

    Personalized federated learning with first order model opti- mization

    Michael Zhang, Karan Sapra, Sanja Fidler, Serena Yeung, and Jose M Alvarez. Personalized federated learning with first order model opti- mization. arXiv preprint arXiv:2012.08565 , 2020

  54. [62]

    Federated learning with non-iid data

    Yue Zhao, Meng Li, Liangzhen Lai, Naveen Suda, Damon Civin, and Vikas Chandra. Federated learning with non-iid data. CoRR, abs/1806.00582, 2018

  55. [63]

    Data-free knowledge distillation for heterogeneous federated learning

    Zhuangdi Zhu, Junyuan Hong, and Jiayu Zhou. Data-free knowledge distillation for heterogeneous federated learning. In International con- ference on machine learning , pages 12878–12889. PMLR, 2021

  56. [2005]

    The record was previously connected to the following departments: Swedish (0101001)

    The information about affiliations in this record was updated in December 2015. The record was previously connected to the following departments: Swedish (0101001)

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.