Pith. sign in

REVIEW 4 major objections 4 minor 74 references

HtFLlib: A Comprehensive Heterogeneous Federated Learning Library and Benchmark

T0 review · 4 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read This paper introduces HtFLlib, a unified library and benchmark that lets heterogeneous federated learning methods be compared across 12 datasets, 40 model architectures, and 10 representative methods.

desk verdict A useful, well-executed HtFL benchmark whose accuracy rankings depend on an ambiguous 'best results' reporting convention that must be clarified before the comparisons are fully trustworthy. read the letter →

arxiv 2506.03954 v1 pith:PWSEWUQN submitted 2025-06-04 cs.LG cs.AIcs.DC

Jianqing Zhang , Xinghao Wu , Yanbing Zhou , Xiaoting Sun , Qiqi Cai , Yang Liu , Yang Hua , Zhenzhe Zheng
show 2 more authors
Jian Cao Qiang Yang
This is my paper
classification cs.LGcs.AIcs.DC
keywords heterogeneousfederatedlearningbenchmarkmodelheterogeneitydataknowledgedistillationprototypesharingmutualpartialparameter
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper argues that heterogeneous federated learning (HtFL), where clients train different model architectures without sharing them, has grown too fast to be evaluated fairly because each research group uses its own datasets, model splits, and implementations. It introduces HtFLlib, an open library and benchmark that standardizes these choices across image, text, and sensor-signal modalities, with 12 datasets, 40 model architectures organized into 19 heterogeneous model groups, and implementations of 10 HtFL methods. The stated goal is to make fair, reproducible comparisons possible and to reveal where each method family, namely partial parameter sharing, mutual distillation, and prototype sharing, does and does not work. If the benchmark is adopted, new HtFL methods can be validated against a common testbed rather than against self-chosen setups.

What carries the argument

The load-bearing machinery is the standardized heterogeneity protocol. Each experiment assigns clients to one of 19 heterogeneous model groups denoted HtFE, HtC, or HtM with a degree $X$, where larger $X$ means greater architectural spread, and forces compatibility by inserting an average pooling layer so every model produces a shared feature dimension $K=512$ before its classifier head. Knowledge transfer is carried by lightweight carriers, namely partial model parameters, a small homogeneous auxiliary model trained by mutual distillation, or class-wise prototypes, depending on the method family. This setup is what makes the comparison fair in principle: same data partitions, same model groups, same feature dimensions, and the same training protocol for all methods.

What would settle it

Re-run the Cifar100 Dirichlet benchmark with $K=64$ and with $K=1024$, and swap the auxiliary model from the smallest to the largest architecture in the group; if relative rankings such as FedTGP over FedKD change materially, the claimed method-neutral calibration fails.

Watch

Extended reading notes

Core claim

The paper's central claim is that HtFLlib is the first heterogeneous federated learning benchmark that unifies data heterogeneity, model heterogeneity, and method implementations in one extensible framework. It defines three data-heterogeneity scenarios (label skew, feature shift, and real-world), four model-heterogeneity families (heterogeneous feature extractors, heterogeneous classifiers, fully heterogeneous models, and per-modality groups), and a modular codebase that re-implements ten representative HtFL methods under identical training steps, batch sizes, and evaluation protocols. The benchmark's headline results show that prototype-sharing methods such as FedTGP lead on image tasks, mutual distillation methods such as FedKD lead on sensor signals, and generator-based knowledge transfer such as FedKTL helps under strong data heterogeneity, but these rankings vary across modalities and heterogeneity levels.

Load-bearing premise

The benchmark assumes its standardizing choices, in particular forcing every client model to output a shared $K=512$ feature vector and choosing the smallest model in each group as the small shared distillation model, do not systematically favor or harm any method family.

Editorial extensions

If this is right

  • A new HtFL method can be positioned against ten existing baselines on a single codebase, making cross-paper accuracy claims directly comparable.
  • Method rankings that hold across image, text, and sensor-signal modalities, such as the strength of prototype sharing on images and its weakness on sensor signals, can guide practitioners toward the right method family for their domain.
  • The reproducible cost tables, in communication megabytes and server and client seconds per round, let researchers see that lightweight prototype methods are not always cheap once server-side training and server-side computation are counted.
  • The black-box pre-trained medical scenario shows that HtFL can improve already-trained heterogeneous models, opening a post-training use case beyond end-to-end federated training.

Reading between the lines

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

  • If HtFLlib becomes the default testbed, the field's implicit overfitting to favorable setups should diminish, because the benchmark fixes data partitions, model groups, and feature dimensions across all comparisons.
  • The benchmark's own results suggest a testable extension: methods that align in logit space degrade less on sensor signals than prototype-alignment methods, so a hybrid carrier transmitting both logits and temporally aware prototypes might combine the strengths of both families.
  • The medical post-training experiment could be extended to map the limits of black-box HtFL, asking at what participation ratio, round budget, and data size per client the collaboration stops improving private pre-trained models.
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 / 4 minor

Summary. This paper introduces HtFLlib, an open-source heterogeneous federated learning (HtFL) library and benchmark. The submission contributes 12 datasets across image, text, and sensor modalities; 19 model-heterogeneity groups covering 40 architectures; 10 implemented HtFL methods organized into partial-parameter-sharing, mutual-distillation, and prototype-sharing categories; and experiments measuring accuracy, convergence, communication cost, and computation cost. The paper's central claim is that HtFLlib is the first comprehensive HtFL benchmark and that its evaluations support conclusions such as the overall strength of FedTGP on image tasks, the weakness of feature-space prototype sharing on text and sensor data, and the robustness of FedMRL to increased model heterogeneity.

Significance. If the quantitative comparisons are trustworthy, HtFLlib is a useful community asset: it unifies dataset/model partitions that were previously scattered across papers, spans modalities and a medical black-box fine-tuning scenario that existing HtFL evaluations ignore, and releases a modular codebase with ten methods. The cost measurements and the explicit discussion of convergence are also valuable. The main risk is that the benchmark's headline rankings are built on an under-specified accuracy-reporting convention and on default calibration choices (K=512, smallest auxiliary model) whose sensitivity is not established. Since the benchmark's purpose is fair and reproducible comparison, these issues are load-bearing rather than cosmetic.

major comments (4)
  1. [Appendix A.1] The sentence 'Each experiment ... is repeated three times ... We report the best results along with error bars' is ambiguous in a way that affects every accuracy table and figure in the paper. If 'best results' means the maximum test accuracy over the 1000 communication rounds, the tables report order statistics of training curves, which can favor methods that spike or oscillate and penalize stable methods, and the 1000-round horizon becomes an arbitrary free parameter. If 'best results' means best-of-three seeds, then error bars computed across seeds do not describe the reported quantity. Please state precisely whether the reported number is final-round accuracy, mean over a stable tail, or best-round accuracy, how the three seeds are combined, and how error bars are computed; if max-over-rounds was used, re-analyze the comparisons with a reporting-convention-insensitive metric or demonstrate that the qualitative rankings are unchanged.
  2. [§4.2 and Fig. 4] The default feature dimension K=512 is fixed in §3.1.2, yet the feature-dimension analysis reports that most methods improve as K grows from 64 to 256, achieve their best at K=256, and that 'a very high K may lead to a performance drop' for FD, FedProto, and FedKTL. Because all main accuracy tables use K=512, this default may systematically disadvantage exactly the prototype-sharing methods that the paper later characterizes as weak on text and sensor data. Please clarify whether K=512 was included in the Fig. 4 sweep, and either choose the default K based on a method-neutral criterion or show that the rankings in Tables 1, 2, 4, 7, and 10 are unchanged at K=256 versus K=512.
  3. [Appendix A.4] Selecting the smallest model in each heterogeneous group as the auxiliary model for FML, FedKD, and FedMRL is a design choice with no sensitivity analysis. Since the auxiliary model's capacity directly affects both the accuracy and the communication cost of mutual-distillation methods, and since the smallest model can be far weaker than the largest in groups such as HtFEimg8 or HtFEtxt6, this choice can bias the category-level comparisons in §4.1.1, §4.3, and §4.4. Please justify this protocol or test at least one alternative auxiliary model to show that the reported category-level conclusions are robust.
  4. [Tables 1 and 4 / §4.1.1] Several categorical claims are made for differences that are within the reported error bars. For example, Table 1 reports FedTGP 46.94±0.12 versus FedKTL 46.94±0.23 on Cifar100 Dirichlet, and Table 4 reports FedTGP 43.17±0.23 versus FedKTL 43.16±0.82 with 50 clients, yet §4.1.1 states that FedTGP 'outperforms all baselines in most cases.' Please add pairwise significance tests or explicitly avoid ranking methods whose differences are smaller than the combined error bars; otherwise the benchmark's qualitative takeaways are not statistically supported.
minor comments (4)
  1. [Table 8] The FD row reads '8 7.73±0.17' and should read '87.73±0.17'.
  2. [§4.1.4 and Table 3] The text says FedProto requires 540, 533, and 570 rounds for α=0.01, 0.1, and 1, but the table lists 540, 533, 570, and 369 for α=0.01, 0.1, 0.5, and 1; please correct the sentence and include α=0.5.
  3. [§3.2.2 and Table 6] The dataset is listed as 'KV ASIR' in Section 3.2.2 and as 'KVASIR' in Table 6; standardize the spelling to 'Kvasir'.
  4. [Figures 4, 6, 7, and 8] These figures show sweeps or training curves without error bars; if error bars were omitted for clarity, state this in the captions and point to where the full numbers are available.

Circularity Check

0 steps flagged · score 2.0 of 10

No circular derivation: HtFLlib's claims are empirical benchmark comparisons; self-citations are non-load-bearing, so the score reflects only minor self-citation, not circular reasoning.

full rationale

The paper makes no mathematical derivation, fitted prediction, or first-principles claim that could reduce to its own inputs. Its contributions are an open-source library, 12 datasets, 40 model architectures, and empirical comparisons of 10 HtFL methods on accuracy, convergence, computation, and communication costs. No equation is defined in terms of the result it is supposed to establish, and no fitted parameter is renamed as a prediction. The authors' own methods FedTGP [62] and FedKTL [63] are included as benchmarked baselines and highlighted as strong performers, and PFLlib [64] is cited as related work; these self-citations are not load-bearing because the benchmark's validity rests on public datasets, stated protocols, and measured results rather than on those papers. The main substantive concerns in the manuscript, such as the ambiguous 'best results' reporting in Appendix A.1 and the design choices of K=512 and smallest-model auxiliary in Sec. 3.1.2 and Appendix A.4, affect comparability and reproducibility but are not circularity: the reported rankings are empirical outcomes of the chosen protocol, not consequences of definitions or self-citations. Score 2 reflects only the presence of minor, non-load-bearing self-citation, not circular reasoning.

Assumptions & free parameters 7 free parameters · 4 assumptions · 0 invented entities

The benchmark's conclusions rest on several hand-chosen hyperparameters and design choices (Dirichlet alpha, feature dimension K, number of clients, etc.) plus domain assumptions about representativeness of datasets and model groups. These are not fitted to data, but they shape all reported comparisons.

free parameters (7)
  • Dirichlet concentration alpha = 0.1
    Controls label skew in the Dirichlet setting. Main results use alpha=0.1; Table 3 shows rankings change with alpha.
  • Feature dimension K = 512
    Enforced by average pooling before classifier heads for compatibility. Figure 4 shows performance varies with K.
  • Number of clients = 20
    Default full-participation scenario. Tables 4 and 6 vary the number of clients.
  • Local training epochs E = 1
    Standard FL setting in prior work. Figure 3 shows some methods are sensitive to E.
  • Batch size = 10
    Used for all experiments, as stated in Appendix A.1.
  • Client learning rate = 0.01
    Used for all experiments, as stated in Appendix A.1.
  • Communication rounds = 1000 for image/text, 500 for sensor
    Maximum number of rounds; convergence conclusions depend on these budgets.
assumptions (4)
  • domain assumption The 12 datasets and 19 model groups are representative of real-world heterogeneous federated learning deployments.
    The comprehensiveness claim rests on this. If the selection is skewed, conclusions about method performance may not generalize.
  • domain assumption The three method categories (partial parameter sharing, mutual distillation, prototype sharing) cover the main design space of HtFL.
    Used to select the 10 methods. A method outside these categories would be absent from the benchmark.
  • domain assumption Test accuracy on a global test set is a valid primary measure of HtFL quality.
    The paper evaluates generalization this way; it does not consider fairness, privacy, or other objectives.
  • domain assumption Standard model architectures and hyperparameters from prior FL work are appropriate for comparing HtFL methods.
    The paper adopts practices from McMahan et al. [35]. If these are biased toward certain methods, rankings could change.

how reviews work

0 comments
Cite this review

Pith. "Pith review of HtFLlib: A Comprehensive Heterogeneous Federated Learning Library and Benchmark." pith.science (2026). https://pith.science/paper/PWSEWUQN

@misc{pith2026250603954,
  author       = {Pith},
  title        = {Pith review of: HtFLlib: A Comprehensive Heterogeneous Federated Learning Library and Benchmark},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PWSEWUQN}},
  note         = {Machine review of arXiv:2506.03954}
}
read the original abstract

As AI evolves, collaboration among heterogeneous models helps overcome data scarcity by enabling knowledge transfer across institutions and devices. Traditional Federated Learning (FL) only supports homogeneous models, limiting collaboration among clients with heterogeneous model architectures. To address this, Heterogeneous Federated Learning (HtFL) methods are developed to enable collaboration across diverse heterogeneous models while tackling the data heterogeneity issue at the same time. However, a comprehensive benchmark for standardized evaluation and analysis of the rapidly growing HtFL methods is lacking. Firstly, the highly varied datasets, model heterogeneity scenarios, and different method implementations become hurdles to making easy and fair comparisons among HtFL methods. Secondly, the effectiveness and robustness of HtFL methods are under-explored in various scenarios, such as the medical domain and sensor signal modality. To fill this gap, we introduce the first Heterogeneous Federated Learning Library (HtFLlib), an easy-to-use and extensible framework that integrates multiple datasets and model heterogeneity scenarios, offering a robust benchmark for research and practical applications. Specifically, HtFLlib integrates (1) 12 datasets spanning various domains, modalities, and data heterogeneity scenarios; (2) 40 model architectures, ranging from small to large, across three modalities; (3) a modularized and easy-to-extend HtFL codebase with implementations of 10 representative HtFL methods; and (4) systematic evaluations in terms of accuracy, convergence, computation costs, and communication costs. We emphasize the advantages and potential of state-of-the-art HtFL methods and hope that HtFLlib will catalyze advancing HtFL research and enable its broader applications. The code is released at https://github.com/TsingZ0/HtFLlib.

Figures

Figures reproduced from arXiv: 2506.03954 by the authors.

Figure 1
Figure 1. Overview of HtFLlib along with experimental results for representative HtFL methods across various heterogeneous model groups, modalities, and data scenarios. Left: Lightweight knowledge carriers are exchanged between the server and clients for knowledge transfer, as sharing entire models is infeasible. Right: Results indicate that methods like FD consistently perform well, while others like FedTGP demonstrate super… view at source ↗
Figure 2
Figure 2. Test accuracy (%) on DomainNet under the feature [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Test accuracy (%) on Cifar100 in the Dirichlet setting [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Test accuracy (%) on Cifar100 in the Dirichlet setting [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: Test accuracy (%) per client on the real-world Came [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: (1) Mutual distillation demonstrates the fastest convergence [PITH_FULL_IMAGE:figures/full_fig_p011_6.png]
Figure 7
Figure 7. Figure 7: The test accuracy curves on the HAR dataset under [PITH_FULL_IMAGE:figures/full_fig_p012_7.png]
Figure 8
Figure 8. Figure 8: The test accuracy curves on the PAMAP2 dataset [PITH_FULL_IMAGE:figures/full_fig_p012_8.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

74 extracted references · 52 canonical work pages

  1. [1]

    Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Floren- cia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, et al. 2023. Gpt-4 technical report.arXiv preprint arXiv:2303.08774 (2023)

    arXiv 2023

  2. [2]

    Md Zahangir Alom, Tarek M Taha, Chris Yakopcic, Stefan Westberg, Paheding Sidike, Mst Shamima Nasrin, Mahmudul Hasan, Brian C Van Essen, Abdul AS Awwal, and Vijayan K Asari. 2019. A state-of-the-art survey on deep learning theory and architectures.electronics8, 3 (2019), 292

    work page 2019

  3. [3]

    Davide Anguita, Alessandro Ghio, Luca Oneto, Xavier Parra, and Jorge L Reyes- Ortiz. 2012. Human activity recognition on smartphones using a multiclass hardware-friendly support vector machine. InAmbient Assisted Living and Home Care: 4th International Workshop, IW AAL 2012, Vitoria-Gasteiz, Spain, December 3-5, 2012. Proceedings 4. Springer, 216–223

    work page 2012

  4. [4]

    Daniel J Beutel, Taner Topal, Akhil Mathur, Xinchi Qiu, Javier Fernandez-Marques, Yan Gao, Lorenzo Sani, Kwing Hei Li, Titouan Parcollet, Pedro Porto Buarque de Gusmão, et al. 2020. Flower: A friendly federated learning research framework. arXiv preprint arXiv:2007.14390(2020)

    arXiv 2020

  5. [5]

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

    arXiv 2018

  6. [6]

    Longbing Cao. 2022. Ai in finance: challenges, techniques, and opportunities. ACM Computing Surveys (CSUR)55, 3 (2022), 1–38

    work page 2022

  7. [7]

    Isabella Castiglioni, Leonardo Rundo, Marina Codari, Giovanni Di Leo, Chris- tian Salvatore, Matteo Interlenghi, Francesca Gallivanone, Andrea Cozzi, Natascha Claudia D’Amico, and Francesco Sardanelli. 2021. AI applications to medical images: From machine learning to deep learning.Physica medica83 (2021), 9–24

    work page 2021

  8. [8]

    Di Chai, Leye Wang, Liu Yang, Junxue Zhang, Kai Chen, and Qiang Yang. 2020. FedEval: A Holistic Evaluation Framework for Federated Learning.arXiv preprint arXiv:2011.09655(2020)

    arXiv 2020

Show all 74 references
  1. [9]

    Kyunghyun Cho, Bart Van Merriënboer, Caglar Gulcehre, Dzmitry Bahdanau, Fethi Bougares, Holger Schwenk, and Yoshua Bengio. 2014. Learning phrase representations using RNN encoder-decoder for statistical machine translation. arXiv preprint arXiv:1406.1078(2014)

    2014 arXiv

  2. [10]

    Patryk Chrabaszcz, Ilya Loshchilov, and Frank Hutter. 2017. A Downsampled Variant of Imagenet as an Alternative to the Cifar Datasets.arXiv preprint arXiv:1707.08819(2017)

    2017 arXiv

  3. [11]

    Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xi- aohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, et al . 2020. An Image is Worth 16x16 Words: Trans- formers for Image Recognition at Scale. InInterna...

    2020

  4. [12]

    Wentao Gao, Omid Tavallaie, Shuaijun Chen, and Albert Zomaya. 2024. Federated learning as a service for hierarchical edge networks with heterogeneous models. InInternational Conference on Service-Oriented Computing. Springer

    2024

  5. [13]

    Yifan Gao, Wei Xia, Dingdu Hu, Wenkui Wang, and Xin Gao. 2024. Desam: De- coupled segment anything model for generalizable medical image segmentation. InInternational Conference on Medical Image Computing and Computer-Assisted Intervention

    2024

  6. [14]

    Hao Guan and Mingxia Liu. 2021. Domain adaptation for medical image analysis: a survey.IEEE Transactions on Biomedical Engineering69, 3 (2021), 1173–1185

    2021

  7. [15]

    Zhiwei Hao, Jianyuan Guo, Kai Han, Yehui Tang, Han Hu, Yunhe Wang, and Chang Xu. 2024. One-for-all: Bridge the gap between heterogeneous architectures in knowledge distillation.Advances in Neural Information Processing Systems (NeurIPS)(2024)

    2024

  8. [16]

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

    2020 arXiv

  9. [17]

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. 2016. Deep Residual Learning for Image Recognition. InIEEE Conference on Computer Vision and Pattern Recognition (CVPR)

    2016

  10. [18]

    S Hochreiter. 1997. Long Short-term Memory.Neural Computation MIT-Press (1997)

    1997

  11. [19]

    Sixu Hu, Yuan Li, Xu Liu, Qinbin Li, Zhaomin Wu, and Bingsheng He. 2022. The oarf benchmark suite: Characterization and implications for federated learning systems.ACM Transactions on Intelligent Systems and Technology (TIST)13, 4 (2022), 1–32

    2022

  12. [20]

    Huimin Huang, Lanfen Lin, Ruofeng Tong, Hongjie Hu, Qiaowei Zhang, Yutaro Iwamoto, Xianhua Han, Yen-Wei Chen, and Jian Wu. 2020. Unet 3+: A full- scale connected unet for medical image segmentation. InICASSP 2020-2020 IEEE international conference on acoustics, speech and sign...

    2020

  13. [21]

    Eunjeong Jeong, Seungeun Oh, Hyesung Kim, Jihong Park, Mehdi Bennis, and Seong-Lyun Kim. 2018. Communication-efficient on-device machine learning: Federated distillation and augmentation under non-iid private data.arXiv preprint arXiv:1811.11479(2018)

    2018 arXiv

  14. [22]

    Armand Joulin, Edouard Grave, Piotr Bojanowski, and Tomas Mikolov. 2016. Bag of tricks for efficient text classification.arXiv preprint arXiv:1607.01759(2016). KDD ’25, August 3–7, 2025, Toronto, ON, Canada Jianqing Zhang et al

    2016 arXiv

  15. [23]

    Peter Kairouz, H Brendan McMahan, Brendan Avent, Aurélien Bellet, Mehdi Bennis, Arjun Nitin Bhagoji, Kallista Bonawitz, Zachary Charles, Graham Cor- mode, Rachel Cummings, et al. 2019. Advances and Open Problems in Federated Learning.arXiv preprint arXiv:1912.04977(2019)

    2019 arXiv

  16. [24]

    Ruhul Amin Khalil, Nasir Saeed, Mudassir Masood, Yasaman Moradi Fard, Mohamed-Slim Alouini, and Tareq Y Al-Naffouri. 2021. Deep learning in the industrial internet of things: Potentials, challenges, and emerging applications. IEEE Internet of Things Journal8, 14 (2021), 11016–11040

    2021

  17. [25]

    David G Kleinbaum, K Dietz, M Gail, Mitchel Klein, and Mitchell Klein. 2002. Logistic regression. Springer

    2002

  18. [26]

    Pang Wei Koh, Shiori Sagawa, Henrik Marklund, Sang Michael Xie, Marvin Zhang, Akshay Balsubramani, Weihua Hu, Michihiro Yasunaga, Richard Lanas Phillips, Irena Gao, et al. 2021. Wilds: A benchmark of in-the-wild distribution shifts. InInternational Conference on Machine Learni...

    2021

  19. [27]

    Alex Krizhevsky and Hinton Geoffrey. 2009. Learning Multiple Layers of Features From Tiny Images.Technical Report(2009)

    2009

  20. [28]

    Yann LeCun, Léon Bottou, Yoshua Bengio, and Patrick Haffner. 1998. Gradient- Based Learning Applied to Document Recognition.Proc. IEEE86, 11 (1998), 2278–2324

    1998

  21. [29]

    Qinbin Li, Yiqun Diao, Quan Chen, and Bingsheng He. 2022. Federated learning on non-iid data silos: An experimental study. In2022 IEEE 38th international conference on data engineering (ICDE). IEEE

    2022

  22. [30]

    Paul Pu Liang, Terrance Liu, Liu Ziyin, Nicholas B Allen, Randy P Auerbach, David Brent, Ruslan Salakhutdinov, and Louis-Philippe Morency. 2020. Think locally, act globally: Federated learning with local and global representations. arXiv preprint arXiv:2001.01523(2020)

    2020 arXiv

  23. [31]

    Tao Lin, Lingjing Kong, Sebastian U Stich, and Martin Jaggi. 2020. Ensemble distillation for robust model fusion in federated learning.Advances in Neural Information Processing Systems (NeurIPS)33 (2020), 2351–2363

    2020

  24. [32]

    Aixin Liu, Bei Feng, Bing Xue, Bingxuan Wang, Bochao Wu, Chengda Lu, Cheng- gang Zhao, Chengqi Deng, Chenyu Zhang, Chong Ruan, et al. 2024. Deepseek-v3 technical report.arXiv preprint arXiv:2412.19437(2024)

    2024 arXiv

  25. [33]

    Yang Liu, Tao Fan, Tianjian Chen, Qian Xu, and Qiang Yang. 2021. Fate: An industrial grade platform for collaborative learning with data protection.The Journal of Machine Learning Research22, 1 (2021), 10320–10325

    2021

  26. [34]

    Jun Ma, Yuting He, Feifei Li, Lin Han, Chenyu You, and Bo Wang. 2024. Segment anything in medical images.Nature Communications15, 1 (2024), 654

    2024

  27. [35]

    Brendan McMahan, Eider Moore, Daniel Ramage, Seth Hampson, and Blaise Aguera y Arcas. 2017. Communication-Efficient Learning of Deep Net- works from Decentralized Data. InInternational Conference on Artificial Intelli- gence and Statistics (AISTATS)

    2017

  28. [36]

    Dinh C Nguyen, Ming Ding, Pubudu N Pathirana, Aruna Seneviratne, Jun Li, and H Vincent Poor. 2021. Federated Learning for Internet of Things: A Comprehen- sive Survey.IEEE Communications Surveys & Tutorials23, 3 (2021), 1622–1658

    2021

  29. [37]

    Maria-Elena Nilsback and Andrew Zisserman. 2008. Automated flower classifica- tion over a large number of classes. In2008 Sixth Indian conference on computer vision, graphics & image processing. IEEE, 722–729

    2008

  30. [38]

    Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, Zeming Lin, Natalia Gimelshein, Luca Antiga, et al. 2019. Pytorch: An imperative style, high-performance deep learning library.Advances in Neural Information Processing Systems...

    2019

  31. [39]

    Xingchao Peng, Qinxun Bai, Xide Xia, Zijun Huang, Kate Saenko, and Bo Wang

  32. [40]

    Konstantin Pogorelov, Kristin Ranheim Randel, Carsten Griwodz, Sigrun Losada Eskeland, Thomas de Lange, Dag Johansen, Concetto Spampinato, Duc-Tien Dang-Nguyen, Mathias Lux, Peter Thelin Schmidt, et al. 2017. Kvasir: A multi- class image dataset for computer aided gastrointest...

    2017

  33. [41]

    Attila Reiss and Didier Stricker. 2012. Introducing a new benchmarked dataset for activity monitoring. In2012 16th international symposium on wearable computers

    2012

  34. [42]

    Mark Sandler, Andrew Howard, Menglong Zhu, Andrey Zhmoginov, and Liang- Chieh Chen. 2018. Mobilenetv2: Inverted residuals and linear bottlenecks. In IEEE Conference on Computer Vision and Pattern Recognition (CVPR)

    2018

  35. [43]

    Iqbal H Sarker. 2021. Deep learning: a comprehensive overview on techniques, taxonomy, applications and research directions.SN computer science2, 6 (2021), 1–20

    2021

  36. [44]

    Mike Schuster and Kuldip K Paliwal. 1997. Bidirectional recurrent neural net- works.IEEE transactions on Signal Processing45, 11 (1997), 2673–2681

    1997

  37. [45]

    Tao Shen, Jie Zhang, Xinkang Jia, Fengda Zhang, Gang Huang, Pan Zhou, Kun Kuang, Fei Wu, and Chao Wu. 2020. Federated mutual learning.arXiv preprint arXiv:2006.16765(2020)

    2020 arXiv

  38. [46]

    Christian Szegedy, Wei Liu, Yangqing Jia, Pierre Sermanet, Scott Reed, Dragomir Anguelov, Dumitru Erhan, Vincent Vanhoucke, and Andrew Rabinovich. 2015. Going deeper with convolutions. InIEEE Conference on Computer Vision and Pattern Recognition (CVPR)

    2015

  39. [47]

    Yue Tan, Guodong Long, Lu Liu, Tianyi Zhou, Qinghua Lu, Jing Jiang, and Chengqi Zhang. 2022. Fedproto: Federated Prototype Learning across Heteroge- neous Clients. InAAAI Conference on Artificial Intelligence (AAAI)

    2022

  40. [48]

    Hugo Touvron, Louis Martin, Kevin Stone, Peter Albert, Amjad Almahairi, Yas- mine Babaei, Nikolay Bashlykov, Soumya Batra, Prajjwal Bhargava, Shruti Bhos- ale, et al. 2023. Llama 2: Open foundation and fine-tuned chat models.arXiv preprint arXiv:2307.09288(2023)

    2023 arXiv

  41. [49]

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. 2017. Attention is All You Need. InAdvances in Neural Information Processing Systems (NeurIPS)

    2017

  42. [50]

    Linda Wang, Zhong Qiu Lin, and Alexander Wong. 2020. Covid-net: A tailored deep convolutional neural network design for detection of covid-19 cases from chest x-ray images.Scientific reports10, 1 (2020), 19549

    2020

  43. [51]

    Wenqi Wei and Ling Liu. 2025. Trustworthy distributed ai systems: Robustness, privacy, and governance.Comput. Surveys57, 6 (2025), 1–42

    2025

  44. [52]

    Chuhan Wu, Fangzhao Wu, Lingjuan Lyu, Yongfeng Huang, and Xing Xie. 2022. Communication-efficient federated learning via knowledge distillation.Nature communications13, 1 (2022), 2032

    2022

  45. [53]

    Shanshan Wu, Tian Li, Zachary Charles, Yu Xiao, Ken Liu, Zheng Xu, and Vir- ginia Smith. 2022. Motley: Benchmarking Heterogeneity and Personalization in Federated Learning. InWorkshop on Federated Learning: Recent Advances and New Challenges (in Conjunction with NeurIPS 2022)

    2022

  46. [54]

    Yuexiang Xie, Zhen Wang, Dawei Gao, Daoyuan Chen, Liuyi Yao, Weirui Kuang, Yaliang Li, Bolin Ding, and Jingren Zhou. 2023. FederatedScope: A Flexible Fed- erated Learning Platform for Heterogeneity.Proceedings of the VLDB Endowment 16, 5 (2023), 1059–1072

    2023

  47. [55]

    Qiang Yang, Yang Liu, Tianjian Chen, and Yongxin Tong. 2019. Federated Machine Learning: Concept and Applications.ACM Transactions on Intelligent Systems and Technology10, 2 (2019), 1–19

    2019

  48. [56]

    Mang Ye, Xiuwen Fang, Bo Du, Pong C Yuen, and Dacheng Tao. 2023. Hetero- geneous federated learning: State-of-the-art and research challenges.Comput. Surveys56, 3 (2023), 1–44

    2023

  49. [57]

    Liping Yi, Gang Wang, Xiaoguang Liu, Zhuan Shi, and Han Yu. 2023. FedGH: Het- erogeneous Federated Learning with Generalized Global Header. InProceedings of the 31st ACM International Conference on Multimedia

    2023

  50. [58]

    Liping Yi, Han Yu, Chao Ren, Gang Wang, Xiaoguang Liu, and Xiaoxiao Li. 2024. Federated Model Heterogeneous Matryoshka Representation Learning.arXiv preprint arXiv:2406.00488(2024)

    2024 arXiv

  51. [59]

    Dun Zeng, Siqi Liang, Xiangjing Hu, Hui Wang, and Zenglin Xu. 2023. FedLab: A Flexible Federated Learning Framework.J. Mach. Learn. Res.24 (2023), 100–1

    2023

  52. [60]

    Ming Zeng, Le T Nguyen, Bo Yu, Ole J Mengshoel, Jiang Zhu, Pang Wu, and Joy Zhang. 2014. Convolutional neural networks for human activity recognition using mobile sensors. In6th international conference on mobile computing, applications and services. IEEE, 197–205

    2014

  53. [61]

    Jianqing Zhang, Yang Hua, Hao Wang, Tao Song, Zhengui Xue, Ruhui Ma, and Haibing Guan. 2023. FedALA: Adaptive Local Aggregation for Personalized Federated Learning. InAAAI Conference on Artificial Intelligence (AAAI)

    2023

  54. [62]

    Jianqing Zhang, Yang Liu, Yang Hua, and Jian Cao. 2024. Fedtgp: Trainable global prototypes with adaptive-margin-enhanced contrastive learning for data and model heterogeneity in federated learning. InAAAI Conference on Artificial Intelligence (AAAI)

    2024

  55. [63]

    Jianqing Zhang, Yang Liu, Yang Hua, and Jian Cao. 2024. An upload-efficient scheme for transferring knowledge from a server-side pre-trained generator to clients in heterogeneous federated learning. InIEEE Conference on Computer Vision and Pattern Recognition (CVPR)

    2024

  56. [64]

    Jianqing Zhang, Yang Liu, Yang Hua, Hao Wang, Tao Song, Zhengui Xue, Ruhui Ma, and Jian Cao. 2025. PFLlib: A Beginner-Friendly and Comprehensive Person- alized Federated Learning Library and Benchmark.Journal of Machine Learning Research26, 50 (2025), 1–10

    2025

  57. [65]

    Sixin Zhang, Anna E Choromanska, and Yann LeCun. 2015. Deep Learning with Elastic Averaging SGD.Advances in Neural Information Processing Systems (NeurIPS)(2015)

    2015

  58. [66]

    Xiang Zhang, Junbo Zhao, and Yann LeCun. 2015. Character-Level Convolutional Networks for Text Classification. InAdvances in Neural Information Processing Systems (NeurIPS)

    2015

  59. [67]

    Ying Zhang, Tao Xiang, Timothy M Hospedales, and Huchuan Lu. 2018. Deep mutual learning. InIEEE Conference on Computer Vision and Pattern Recognition (CVPR)

    2018

  60. [68]

    Zilong Zhong, Jonathan Li, Lingfei Ma, Han Jiang, and He Zhao. 2017. Deep resid- ual networks for hyperspectral image classification. In2017 IEEE International Geoscience and Remote Sensing Symposium (IGARSS). IEEE, 1824–1827

    2017

  61. [69]

    Pan Zhou, Jiashi Feng, Chao Ma, Caiming Xiong, Steven Chu Hong Hoi, et al

  62. [70]

    Zhuangdi Zhu, Junyuan Hong, and Jiayu Zhou. 2021. Data-Free Knowledge Distillation for Heterogeneous Federated Learning. InInternational Conference on Machine Learning (ICML). HtFLlib: A Comprehensive Heterogeneous Federated Learning Library and Benchmark KDD ’25, August 3–7, ...

    2021

  63. [73]

    (1) fastText: This model uses an embedding layer followed by a linear hidden layer and a final output layer

    and Logistic Regression [25]. (1) fastText: This model uses an embedding layer followed by a linear hidden layer and a final output layer. (2) Logistic Regression: This is a traditional linear classifier applied directly to the word embeddings. • Architectures in HtFE𝑡𝑥𝑡 4: Th...

  64. [74]

    (1) HARCNN: The model consists of 2 convolutional layers, followed by 2 pooling layers and 3 fully connected layers

    with varying strides (1, 2). (1) HARCNN: The model consists of 2 convolutional layers, followed by 2 pooling layers and 3 fully connected layers. (2) The strides of the convolutional layers are set to 1 and 2. • Architectures in HtFE 𝑠𝑒𝑛 3: This model group is similar to HtFE𝑠...

    2025

  65. [2019]

    InIEEE Interna- tional Conference on Computer Vision (ICCV)

    Moment Matching for Multi-Source Domain Adaptation. InIEEE Interna- tional Conference on Computer Vision (ICCV)

  66. [2020]

    Towards theoretically understanding why sgd generalizes better than adam in deep learning.Advances in Neural Information Processing Systems (NeurIPS) (2020)

    2020

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.