Personalized One-shot Federated Graph Learning for Heterogeneous Clients

Guochen Yan; Luyuan Xie; Qingni Shen; Xunkai Li; Yuejian Fang; Zhonghai Wu

REVIEW 3 major objections 6 minor 1 cited by

Personalized One-shot Federated Graph Learning for Heterogeneous Clients

T0 review · 3 major / 6 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read One-shot federated graph learning beats 100-round methods on average, the paper claims.

desk verdict First real one-shot personalized FGL paper with broad, consistent empirical gains, but the headline unbiasedness claim for its variance aggregation is wrong and needs fixing before the method's guarantees can be taken at face value. read the letter →

arxiv 2411.11304 v7 pith:CQ354PE7 submitted 2024-11-18 cs.LG

Guochen Yan , Xunkai Li , Luyuan Xie , Qingni Shen , Yuejian Fang , Zhonghai Wu This is my paper · ORCID

classification cs.LG

keywords one-shotfederatedlearninggraphpersonalizednodeclassificationsurrogateknowledgedistillationlabelpropagationdataheterogeneity

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper tries to establish that personalized federated learning on graphs can be done in a single communication round without sacrificing accuracy or fairness. The proposed method, O-pFGL, has each client estimate class-wise means and variances of propagated node features, the server combines these statistics into a small global surrogate graph, and each client then trains its own model in two stages with a distillation step that protects minority classes. The claim is that this one-shot procedure outperforms existing federated graph learning and one-shot federated learning baselines across 14 datasets, and even beats multi-round methods given 100 rounds. If true, it would make collaborative graph learning practical where communication is expensive or risky, and would let clients keep both their data and their model architectures private.

What carries the argument

The load-bearing object is the global surrogate graph $G' = (A', X', Y')$, a small graph (under 300 nodes in the experiments) generated on the server so that its class-wise propagated feature means and variances match the unbiased aggregated global statistics. The adjacency matrix $A'$ comes from a learnable link predictor $g_\theta$ over node features, with a smoothness loss $\mathcal{L}_{\text{smt}}$ to encourage a realistic structure, and the node features $X'$ are optimized against the alignment loss $\mathcal{L}_{\text{align}}$. Around this object sit two supporting mechanisms: the Homophily-guided Reliable node Expansion (HRE) strategy, which uses label propagation and class homophily to expand scarce labeled sets for more accurate statistics, and node-adaptive distillation, which weights the KL divergence from the global teacher by $\gamma_i = \beta \tilde{y}_i \cdot w_{\text{dist}}$ with $w_{\text{dist}}[c] = 1/(1 + \log(H(c) + 1))$, so that nodes in low-homophily or minority classes inherit more global knowledge.

What would settle it

Run O-pFGL on a synthetic heterophilic graph whose class-wise feature means and variances are constructed to be identical across classes while labels are determined by triadic closure or other higher-order patterns. If the method fails to beat standalone local training on that graph, the claim that class-wise first and second moments of propagated features suffice for a useful surrogate graph is refuted.

Watch

Extended reading notes

Core claim

The central discovery is that the information needed to personalize a graph model across heterogeneous clients can be carried by class-wise first and second moments of propagated features, rather than by model parameters or raw graph data. From these aggregated moments the server synthesizes a small surrogate graph whose propagated feature distribution matches the recovered global distribution, and this surrogate graph serves as a teacher for a two-stage local training procedure. In the second stage, a node-adaptive distillation weight, derived from class homophily and soft labels, decides per node how much to trust the global teacher versus local fine-tuning, which the paper argues corrects the bias of fine-tuning toward majority classes. The paper reports consistent gains in accuracy and F1-macro over eleven baselines on 14 real-world datasets, and better average performance than multi-round methods allowed 100 communication rounds.

Load-bearing premise

The method works only if a small graph whose class-wise propagated feature means and variances match the global statistics captures enough of the true graph structure and feature distribution to train a teacher model that improves each client's personalization.

Editorial extensions

If this is right

Communication cost becomes independent of model size: clients upload only $O(C \cdot h \cdot f)$ statistics and download a small surrogate graph, which scales better than parameter exchange as models grow.
Because no model parameters are shared, clients can use heterogeneous architectures and protect model intellectual property, with statistics upload compatible with Secure Aggregation protocols.
The node-adaptive distillation is claimed to raise F1-macro substantially over fine-tuning, implying better performance on minority classes in imbalanced local graphs.
The method reportedly beats multi-round federated baselines trained for 100 rounds on average, suggesting one-shot collaboration may be sufficient for node classification in many real-world regimes.
The approach carries over to inductive datasets and other GNN backbones (GraphSage, GAT, SGC) per the experiments.

Reading between the lines

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

The method's success would imply that cross-client graph knowledge transfer can be reduced to moment matching of propagated features, a much weaker commitment than sharing structure; this suggests a general design principle for one-shot federated learning on non-Euclidean data.
A testable extension is to apply the surrogate-graph idea to graph-level tasks (graph classification) or to link prediction, where the same statistics-plus-smoothness recipe might need higher-order moments to capture motif structure.
The reliance on class-wise moments predicts a failure mode: on graphs where classes are separated by higher-order structural patterns (e.g., role detection in heterophilic networks), the surrogate graph may be information-poor; experiments that withhold second-order moments or scramble topology would isolate how much structure the method actually uses.
Because the method assumes honest clients and an honest server, an adversary-free setting is load-bearing for the privacy claims; a malicious server could still infer class proportions and feature statistics from the aggregated moments.

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.

Desk Editor's Note

read the letter

Short version: this is the first paper I know that combines one-shot communication with personalization for federated node classification. The empirical sweep is unusually wide—14 datasets, Louvain/Metis partitions, 10/20/1000 clients, homophilic/heterophilic, inductive, heterogeneous models, several GNN backbones—and the gains over the next-best baselines, especially in F1-macro, are consistent. The two-stage node-adaptive distillation is a sensible mechanism: train a global teacher on a server-side surrogate graph, then fine-tune locally with per-node KL weighting based on class homophily. HRE is a reasonable plug-in. The evaluation is on held-out nodes, so the result is not a circular fit. That part holds up.

The soft spot is real and load-bearing. Eq. 8 divides the total sum of squares by N_c - m and calls the result an unbiased global variance. That is not unbiased. The unbiased estimator of the total variance is S_total/(N_c - 1); dividing by N_c - m biases it upward by (N_c-1)/(N_c-m), and when a minority class has fewer total samples than there are clients, the denominator is non-positive. This is exactly the target the surrogate graph aligns to in Eq. 10, so the recovered global distribution is distorted for the minority classes the paper most wants to protect. The claim in Eq. 8 and repeated in Appendix J that the recovered statistics are unbiased regardless of non-IIDness is simply false.

Does that kill the empirical result? Not automatically—the reported gains could still hold because the optimization relaxes the alignment loss, and the method may work fine with a biased variance target. But it removes the paper's main theoretical guarantee, and the method is otherwise built on an unproven sufficiency assumption: that class-wise means and variances of propagated features capture enough of the global graph structure. That assumption is not justified.

Also: no code is released, and the per-dataset hyperparameters (p, f_th, d_th, K, beta, delta, alpha) are many and tuned per dataset. The appendix gives sensitivity for some, but an off-the-shelf deployment story is weak.

Who this is for: anyone working on one-shot federated learning or personalized FGL. The problem framing is valuable and the empirical evidence is worth taking seriously. A serious referee should get this paper, but the authors need to correct the variance claim, release code, and either prove or drop the sufficiency assumption. I'd send it out rather than desk-reject.

Referee Report

3 major / 6 minor

Summary. The manuscript proposes O-pFGL, a one-shot personalized federated graph learning method for node classification. Each client estimates class-wise means and variances of propagated node features and uploads these statistics in a form compatible with Secure Aggregation. The server aggregates the statistics into a purported global class-wise distribution and synthesizes a small global surrogate graph by matching the first two moments of propagated features and enforcing a smoothness prior. Clients then train a generalized model on the surrogate graph (Stage 1) and fine-tune it on local data with node-adaptive knowledge distillation that up-weights minority and low-homophily nodes (Stage 2). The method supports heterogeneous client models. Experiments on 14 real-world datasets compare O-pFGL against 11 baselines under Louvain/Metis partitions with 10 and 20 clients, plus heterophilic, inductive, model-heterogeneity, and large-client settings. The paper reports consistent accuracy and F1-macro gains, and Appendix D.2 reports better average performance than multi-round baselines run for 100 rounds.

Significance. The problem is timely and the empirical study is unusually broad: 14 datasets, multiple partition schemes, 10/20 clients, inductive and heterophilic benchmarks, model-heterogeneity experiments, and a detailed ablation study. The evaluation is on held-out nodes and the surrogate graph is generated from training-set statistics, so the reported gains are not fitted values; the empirical contribution is genuine evidence for the approach. The communication cost (class-wise statistics plus a small surrogate graph) and the compatibility with Secure Aggregation are practical strengths. However, the paper's central theoretical claim that the aggregated global variance is unbiased is mathematically incorrect, and the implementation of the variance alignment for surrogate graphs with one node per class is left unspecified. The main result is therefore supported empirically but not by the stated derivation; the manuscript needs a corrected aggregation formula and a corresponding re-examination of the surrogate-generation loss before the claims can be accepted.

major comments (3)

[Global Surrogate Graph Generation, Eq. (8)] The aggregation formula in Eq. (8) is not an unbiased estimator of the global class-wise feature variance, contradicting the sentence immediately after it. Under IID sampling from a common distribution per class, the total sum of squares in the numerator has expectation (N_c - 1) sigma_c^2, so dividing by N_c - m yields E[s^{2c}] = (N_c - 1)/(N_c - m) sigma_c^2, which overestimates the variance. Under heterogeneous client means, the between-client term additionally inflates the estimate, so the bias grows with non-IIDness. Moreover, if a class has N_c < m total samples, which can occur under the label-skew partitions used in the experiments, the denominator is non-positive and the formula yields an undefined or negative variance with no stated fallback. Since Eq. (10) aligns the surrogate graph's variance to this value, the distortion propagates into the global model and the distillation stage. The manuscript should either divide by N_c - 1 to estimate the pooled total variance or explicitly define and justify the target quantity being estimated, and it must specify a valid fallback for small N_c.
[Global Surrogate Graph Generation, Eq. (5) and Appendix C] The surrogate graph generation uses a sample variance defined in Eq. (5) with denominator |V| - 1. Appendix C states that on Cora, CiteSeer, and PubMed the global surrogate graph is set to exactly one node per class. For a class with one node, the sample variance is undefined (division by zero), yet Eq. (10) includes the term (s'^{2c} - s^{2c})^2. The paper does not describe the fallback estimator actually used in the implementation, nor how the variance alignment is computed for one-node classes. This is a load-bearing implementation detail because the entire surrogate graph is generated from these moments. The text must state the actual variance estimator used for the surrogate graph or adjust the node-count configuration so that the stated equations are well defined.
[Global Surrogate Graph Generation, Eqs. (10)-(11)] The surrogate graph is generated by matching only the first two moments of propagated features plus a smoothness prior. The paper does not provide evidence that these statistics capture the structural information needed to transfer global knowledge in heterophilic or structurally heterogeneous graphs, such as degree distributions, edge homophily, or higher-order neighborhood structure. Because the surrogate graph is the only channel through which cross-client knowledge enters Stage 1 and the distillation stage, a diagnostic test would substantiate the mechanism: for example, compare the surrogate graph's edge homophily and degree distribution with the client graphs, or compare the Stage-1 model's accuracy against a model trained on an oracle global graph. Without such evidence, the end-to-end gains demonstrate that the method works but do not establish that the proposed surrogate-graph construction is the reason.

minor comments (6)

[Feature Distribution Estimation, Eq. (5)] The local class-wise variance is undefined when a client has exactly one labeled node of a class, since the denominator |V^c_{k,L}| - 1 is zero. The manuscript says the statistics are estimated only for classes with sufficient labeled nodes but never defines the minimum count or a fallback estimator.
[Table 3] The Actor row lists the property as 'Heterophlic'; this should be 'Heterophilic'.
[Appendix D.2] The sentence 'We set the communication round of these 6 methods to 100' is followed by a list of seven methods (FedAvg, FedPUB, FedGTA, FedTAD, FedSpray, FedLoG, and FedGM); the count should be corrected.
[Preliminaries, Eq. (1)] The definition of hnode(vi) divides by |N_vi ∩ V_L|, which is zero for an isolated labeled node; the paper should state the convention adopted in that case.
[Appendix J] The statement that the aggregation process is 'lossless and unbiased regardless of the data distributions' repeats the incorrect claim from Eq. (8) and should be revised consistently with the corrected variance estimator.
[Figure 2] The text describes the pipeline as comprising five steps, but the figure does not show step indices; adding visible step numbers would make the correspondence between the text and the figure much clearer.

Circularity Check

1 steps flagged · score 6.0 of 10

Partial circularity: HRE moves test-set nodes into training, so part of the reported 'prediction' is a fit to pseudo-labels on the evaluation set.

fitted input called prediction [Preliminaries (test-set definition), Feature Distribution Estimation (Eq. 6-7), Personalization with Node Adaptive Distillation (Eq. 14-15)]
"nodes on the k-th client are partitioned into labeled nodes Vk,L (i.e., training set) and unlabeled nodes Vk,U (i.e., test set). ... Vk,r ={vi|(vi ∈ Vk,U ) ∧ (d(vi) ≥ dth) ∧ (˜yi(c′) ≥ fth) ∧ (c′ ∈ topK(H))}. ... The reliable nodes Vk,r, along with their inferred labels, can expand the original labeled nodes set Vk,L. ... Ldist = Σ_{xi∈Xk} γiLkl(Mk(Ak, xi), MG(Ak, xi))."

By Eq. 6, the reliable nodes Vk,r are explicitly a subset of Vk,U, which the paper defines as the test set. Eq. 7 then uses Vk,r∪Vk,L to compute the class-wise statistics that generate the surrogate graph and train the teacher MG. The Stage-2 objective (Eq. 14-15) further applies a distillation loss over all xi∈Xk, which includes every node of Vk,U. The reported accuracy and F1-macro are computed on Vk,U. Therefore a subset of the test nodes is assigned label-propagation pseudo-labels, used as training targets (directly in Eq. 7 and indirectly through MG), and then counted again as 'predicted' in the final evaluation. The model is fitted to those pseudo-labels, so the test score on that subset is a measure of fit to the pseudo-labeling input, not an independent prediction.

full rationale

The central derivation chain (local class-wise statistics in Eq. 5-7, global aggregation in Eq. 8, surrogate-graph generation in Eq. 9-11, global-teacher training in Eq. 13, and personalized fine-tuning in Eq. 14-15) is not circular in the sense of a parameter being fitted and then renamed as a prediction: the surrogate graph is generated from training-set moments, and the true test labels are never used as training targets. Self-citations to OpenFGL, FedGTA, AdaFGL, and FedTAD are benchmark, baseline, or partition references, not load-bearing uniqueness theorems, so they do not raise the circularity score. The Eq. 8 variance estimator is not actually unbiased for m>1 (a correctness risk, not a circularity). The genuine circular element is evaluative: HRE selects 'reliable' unlabeled nodes from the test set Vk,U, uses their pseudo-labels to augment the training data (Eq. 7), trains the final model also by distillation over all xi∈Xk (Eq. 15), and then reports test metrics on Vk,U. Hence a component of the reported test performance is a fit to pseudo-labels on evaluation nodes rather than an independent prediction, making the SOTA claim partially circular by construction.

Assumptions & free parameters 8 free parameters · 4 assumptions · 0 invented entities

No new physical or independently falsifiable entities are introduced. The global surrogate graph is an algorithmic artifact, not a postulated natural entity. The free parameters are method hyperparameters, and the axioms are the unproven modeling choices on which the method rests.

free parameters (8)

p (surrogate graph size per class, as percentage of class training size) = 0.25%, 0.04%, 1%, 1%, 1%, 0.5%, 0.2%, 0.05%, 0.05%, 2%, 0.1% for ogbn-arxiv, ogbn-products, Computers, Photo, CS…
Tuned per dataset to keep the surrogate graph small while preserving enough class information; affects the amount of global information available for distillation.
f_th (HRE confidence threshold) = tuned within [0.95, 1]
Selects high-confidence soft labels from label propagation for reliable node expansion.
d_th (HRE degree threshold) = tuned according to graph scale
Selects high-degree nodes to reduce variance in pseudo-label estimates.
K (top homophily classes in HRE) = client-defined
Restricts expansion to classes with high class homophily; impacts which pseudo-labeled nodes are added.
beta (distillation scaling) = tuned within [0.1, 1]
Scales the node-adaptive distillation weight gamma_i; too small or too large degrades performance.
alpha (smoothness loss weight) = adopted from Xiao et al. 2024, not reported explicitly
Balances feature alignment and graph smoothness in surrogate generation.
delta (sparsity threshold in link predictor) = hyperparameter, values not reported per dataset
Controls sparsity of the surrogate adjacency matrix; when delta=1 all nodes are isolated.
propagation depth h = h <= 2 in experiments
Number of feature propagation steps in Eq 4; affects the estimated distribution.

assumptions (4)

domain assumption Class-wise mean and variance of propagated features are sufficient to generate a surrogate graph that represents the global graph distribution.
Invoked in 'Global Surrogate Graph Generation' where a small graph is fit to these statistics and treated as a teacher for all clients; no proof that higher-order structure is captured.
ad hoc to paper The aggregation formula in Eq 8 yields an unbiased estimate of global class-wise variance.
The formula divides by N_c - m rather than N_c - 1, so it overestimates the pooled sample variance when m > 1; the unbiasedness claim in the text is inaccurate.
domain assumption Label propagation produces reliable soft labels for high-degree, high-confidence nodes in top-homophily classes.
HRE (Eq 6) assumes these criteria identify nodes whose inferred labels match true labels; if wrong, the estimated statistics are polluted.
domain assumption The global model MG trained on the surrogate graph provides useful knowledge for minority classes during distillation.
Stage 2 distills from MG under the premise that global knowledge counteracts local majority bias; the paper validates this empirically but does not prove it.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Personalized One-shot Federated Graph Learning for Heterogeneous Clients." pith.science (2026). https://pith.science/paper/CQ354PE7

@misc{pith2026241111304,
  author       = {Pith},
  title        = {Pith review of: Personalized One-shot Federated Graph Learning for Heterogeneous Clients},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/CQ354PE7}},
  note         = {Machine review of arXiv:2411.11304}
}

read the original abstract

Federated Graph Learning (FGL) has emerged as a promising paradigm for breaking data silos among distributed private graphs. In practical scenarios involving heterogeneous distributed graph data, personalized Federated Graph Learning (pFGL) aims to enhance model utility by training personalized models tailored to client needs. However, existing pFGL methods often require numerous communication rounds under heterogeneous graphs, leading to significant communication overhead and security concerns. While One-shot Federated Learning (OFL) enables collaboration in a single round, existing OFL methods are designed for image-centric tasks and are ineffective for graph data, leaving a critical gap in the field. Additionally, personalized models derived from existing methods suffer from bias, failing to effectively generalize to the minority. To address these challenges, we propose the first \textbf{O}ne-shot \textbf{p}ersonalized \textbf{F}ederated \textbf{G}raph \textbf{L}earning method (\textbf{O-pFGL}) for node classification, compatible with Secure Aggregation protocols for privacy preservation. Specifically, for effective graph learning in one communication round, our method estimates and aggregates class-wise feature distribution statistics to construct a global surrogate graph on the server, facilitating the training of a global graph model. To mitigate bias, we introduce a two-stage personalized training approach that adaptively balances local personal information and global insights from the surrogate graph, improving both personalization and generalization. Extensive experiments on 14 diverse real-world graph datasets demonstrate that our method significantly outperforms state-of-the-art baselines across various settings.

Figures

Figures reproduced from arXiv: 2411.11304 by the authors.

**Figure 2.** The details are explained in the following sections. [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗

**Figure 2.** The overall pipeline of our proposed method, which comprises 5 steps: (1) Each client estimates the class-wise feature [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

Figures from the paper (9 more)

**Figure 3.** Figure 3: Ablation study under Louvain partition. nodes of major classes, the model directly learns through supervised fine-tuning. For nodes of minor classes, Mk additionally distills the predictive ability from MG. The training objective for the second stage is to minimize: …

**Figure 4.** Figure 4: Communication costs. (3) FedTAD (Zhu et al. 2024b), (4) FedSpray (Fu et al. 2024), (5) FedLoG (Kim et al. 2025), (6) FedGM (Zhang et al. 2025), (7) GHOST (Qian et al. 2025) and 2 OFL methods: (8) DENSE (Zhang et al. 2022), (9) Co-Boost (Dai et al. 2024). We also inclu…

**Figure 5.** Figure 5: Left: Comparison of different personalized training methods. MG is the generalized model trained in Stage 1. Middle: The value of H and wdist for each class. Right: The value of H is determined by both samples quantity and homophily. but they require numerous communica…

**Figure 6.** Figure 6: ), it likely belongs to minor classes or has low homophily. These nodes are often neglected during supervised fine-tuning. Thus γi should be large to incorporate the corresponding global knowledge from MG. b a 𝑤𝑑𝑖𝑠𝑡(𝑐) 𝒚෦𝒂 a b High homophily, prone to be major class …

**Figure 7.** Figure 7: Ablation study under Metis partition. tions are shown in [PITH_FULL_IMAGE:figures/full_fig_p013_7.png]

**Figure 9.** Figure 9: Performance with different β values (C2)). Also, we only select reliable soft labels that potentially belong to the classes with topK high class homophily (i.e., (C3)). However, our method is not sensitive to the selection of the threshold fth, dth, and K values. We …

**Figure 8.** Figure 8: Performance with different K values ciples to choose the number are (1) set the smallest number to capture enough information, (2) keep a similar label distribution with the training set. According to Section C, we pre-set the number of nodes to synthesize to 1 for th…

**Figure 10.** Figure 10: Performance of method with a large number of [PITH_FULL_IMAGE:figures/full_fig_p023_10.png]

**Figure 11.** Figure 11: Performance of method on the subsequent link prediction task on Cora, CiteSeer, PubMed datasets. The x-axis is set [PITH_FULL_IMAGE:figures/full_fig_p024_11.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

A Comprehensive Data-centric Overview of Federated Graph Learning
cs.LG 2025-07 conditional novelty 5.0 of 10

A data-centric taxonomy for Federated Graph Learning that classifies 79 studies by data characteristics and data utilization, plus a discussion of integration with pre-trained large models.

Reference graph

Works this paper leans on

75 extracted references · 35 canonical work pages · cited by 1 Pith paper

[1]

, " * write output.state after.block = add.period write newline

ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.a...
[2]

write newline

" write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...
[3]

Abu-El-Haija, S.; Perozzi, B.; Kapoor, A.; Alipourfard, N.; Lerman, K.; Harutyunyan, H.; Ver Steeg, G.; and Galstyan, A. 2019. Mixhop: Higher-order graph convolutional architectures via sparsified neighborhood mixing. In international conference on machine learning, 21--29. PMLR

work page 2019
[4]

S.; and Conti, M

Acar, A.; Aksu, H.; Uluagac, A. S.; and Conti, M. 2018. A survey on homomorphic encryption schemes: Theory and implementation. ACM Computing Surveys (Csur), 51(4): 1--35

work page 2018
[5]

Aliakbari, J.; \"O stman, J.; et al. 2024. Decoupled Subgraph Federated Learning. arXiv preprint arXiv:2402.19163

work page Pith review arXiv 2024
[6]

Baek, J.; Jeong, W.; Jin, J.; Yoon, J.; and Hwang, S. J. 2023. Personalized subgraph federated learning. In International conference on machine learning, 1396--1415. PMLR

2023
[7]

Bang, D.; Lim, S.; Lee, S.; and Kim, S. 2023. Biomedical knowledge graph learning for drug repurposing by extending guilt-by-association to multiple layers. Nature Communications, 14(1): 3570

2023
[8]

B.; Patel, S.; Ramage, D.; Segal, A.; and Seth, K

Bonawitz, K.; Ivanov, V.; Kreuter, B.; Marcedone, A.; McMahan, H. B.; Patel, S.; Ramage, D.; Segal, A.; and Seth, K. 2017. Practical secure aggregation for privacy-preserving machine learning. In proceedings of the 2017 ACM SIGSAC Conference on Computer and Communications Security, 1175--1191

2017

Show all 75 references

[9]

Chen, C.; Xu, Z.; Hu, W.; Zheng, Z.; and Zhang, J. 2024. FedGL: Federated graph learning framework with global self-supervision. Information Sciences, 657: 119976

2024
[10]

Chen, D.; Lin, Y.; Zhao, G.; Ren, X.; Li, P.; Zhou, J.; and Sun, X. 2021 a . Topology-imbalance learning for semi-supervised node classification. Advances in Neural Information Processing Systems, 34: 29885--29897

2021
[11]

Chen, M.; Zhang, W.; Yuan, Z.; Jia, Y.; and Chen, H. 2021 b . Fede: Embedding knowledge graphs in federated setting. In Proceedings of the 10th International Joint Conference on Knowledge Graphs, 80--88

2021
[12]

Dai, R.; Shen, L.; He, F.; Tian, X.; and Tao, D. 2022. Dispfl: Towards communication-efficient personalized federated learning via decentralized sparse training. In International conference on machine learning, 4587--4604. PMLR

2022
[13]

Dai, R.; Zhang, Y.; Li, A.; Liu, T.; Yang, X.; and Han, B. 2024. Enhancing One-Shot Federated Learning Through Data and Ensemble Co-Boosting. arXiv preprint arXiv:2402.15070

2024 arXiv
[14]

Diao, Y.; Li, Q.; and He, B. 2023. Towards addressing label skews in one-shot federated learning. In The Eleventh International Conference on Learning Representations

2023
[15]

Finn, C.; Abbeel, P.; and Levine, S. 2017. Model-agnostic meta-learning for fast adaptation of deep networks. In International conference on machine learning, 1126--1135. PMLR

2017
[16]

Frasca, F.; Rossi, E.; Eynard, D.; Chamberlain, B.; Bronstein, M.; and Monti, F. 2020. Sign: Scalable inception graph neural networks. arXiv preprint arXiv:2004.11198

2020 arXiv
[17]

Fu, X.; Chen, Z.; He, Y.; Wang, S.; Zhang, B.; Chen, C.; and Li, J. 2025. Virtual Nodes Can Help: Tackling Distribution Shifts in Federated Graph Learning. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 39, 16657--16665

2025
[18]

Fu, X.; Chen, Z.; Zhang, B.; Chen, C.; and Li, J. 2024. Federated graph learning with structure proxy alignment. In Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 827--838

2024
[19]

Fu, X.; Zhang, B.; Dong, Y.; Chen, C.; and Li, J. 2022. Federated graph machine learning: A survey of concepts, techniques, and applications. ACM SIGKDD Explorations Newsletter, 24(2): 32--47

2022
[20]

Gasteiger, J.; Bojchevski, A.; and G \"u nnemann, S. 2018. Predict then propagate: Graph neural networks meet personalized pagerank. arXiv preprint arXiv:1810.05997

2018 arXiv
[21]

Gu, Z.; Zhang, K.; Bai, G.; Chen, L.; Zhao, L.; and Yang, C. 2023. Dynamic activation of clients and parameters for federated learning over heterogeneous graphs. In 2023 IEEE 39th International Conference on Data Engineering (ICDE), 1597--1610. IEEE

2023
[22]

Guha, N.; Talwalkar, A.; and Smith, V. 2019. One-shot federated learning. arXiv preprint arXiv:1902.11175

2019 arXiv
[23]

Hamilton, W.; Ying, Z.; and Leskovec, J. 2017. Inductive representation learning on large graphs. Advances in neural information processing systems, 30

2017
[24]

E.; Luz-Ricca, E.; and Shao, H

Heinbaugh, C. E.; Luz-Ricca, E.; and Shao, H. 2023. Data-free one-shot federated learning under very high statistical heterogeneity. In The Eleventh International Conference on Learning Representations

2023
[25]

W.; Laskey, K

Holland, P. W.; Laskey, K. B.; and Leinhardt, S. 1983. Stochastic blockmodels: First steps. Social networks, 5(2): 109--137

1983
[26]

Hu, W.; Fey, M.; Zitnik, M.; Dong, Y.; Ren, H.; Liu, B.; Catasta, M.; and Leskovec, J. 2020. Open graph benchmark: Datasets for machine learning on graphs. Advances in neural information processing systems, 33: 22118--22133

2020
[27]

Huang, W.; Wan, G.; Ye, M.; and Du, B. 2024. Federated graph semantic and structural learning. arXiv preprint arXiv:2406.18937

2024 arXiv
[28]

Hyun, W.; Lee, J.; and Suh, B. 2023. Anti-money laundering in cryptocurrency via multi-relational graph neural network. In Pacific-Asia Conference on Knowledge Discovery and Data Mining, 118--130. Springer

2023
[29]

Iscen, A.; Tolias, G.; Avrithis, Y.; and Chum, O. 2019. Label propagation for deep semi-supervised learning. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 5070--5079

2019
[30]

A.; Jordan, M

Jacobs, R. A.; Jordan, M. I.; Nowlan, S. J.; and Hinton, G. E. 1991. Adaptive mixtures of local experts. Neural computation, 3(1): 79--87

1991
[31]

B.; Avent, B.; Bellet, A.; Bennis, M.; Bhagoji, A

Kairouz, P.; McMahan, H. B.; Avent, B.; Bellet, A.; Bennis, M.; Bhagoji, A. N.; Bonawitz, K.; Charles, Z.; Cormode, G.; Cummings, R.; et al. 2021. Advances and open problems in federated learning. Foundations and trends in machine learning , 14(1--2): 1--210

2021
[32]

P.; Kale, S.; Mohri, M.; Reddi, S.; Stich, S.; and Suresh, A

Karimireddy, S. P.; Kale, S.; Mohri, M.; Reddi, S.; Stich, S.; and Suresh, A. T. 2020. Scaffold: Stochastic controlled averaging for federated learning. In International conference on machine learning, 5132--5143. PMLR

2020
[33]

Kim, S.; Lee, Y.; Oh, Y.; Lee, N.; Yun, S.; Lee, J.; Kim, S.; Yang, C.; and Park, C. 2025. Subgraph federated learning for local generalization. arXiv preprint arXiv:2503.03995

2025 arXiv
[34]

N.; and Welling, M

Kipf, T. N.; and Welling, M. 2016. Semi-supervised classification with graph convolutional networks. arXiv preprint arXiv:1609.02907

2016 arXiv
[35]

Li, N. 2010. Research on Diffie-Hellman key exchange protocol. In 2010 2nd International Conference on Computer Engineering and Technology, volume 4, V4--634. IEEE

2010
[36]

Li, X.; Wu, Z.; Zhang, W.; Sun, H.; Li, R.-H.; and Wang, G. 2024 a . AdaFGL: A New Paradigm for Federated Node Classification with Topology Heterogeneity. arXiv preprint arXiv:2401.11750

2024 arXiv
[37]

Li, X.; Wu, Z.; Zhang, W.; Zhu, Y.; Li, R.-H.; and Wang, G. 2024 b . Fedgta: Topology-aware averaging for federated graph learning. arXiv preprint arXiv:2401.11755

2024 arXiv
[38]

Li, X.; Zhu, Y.; Pang, B.; Yan, G.; Yan, Y.; Li, Z.; Wu, Z.; Zhang, W.; Li, R.-H.; and Wang, G. 2024 c . OpenFGL: A Comprehensive Benchmarks for Federated Graph Learning. arXiv preprint arXiv:2408.16288

2024 arXiv
[39]

Luan, S.; Hua, C.; Lu, Q.; Ma, L.; Wu, L.; Wang, X.; Xu, M.; Chang, X.-W.; Precup, D.; Ying, R.; et al. 2024. The heterophilic graph learning handbook: Benchmarks, models, theoretical analysis, applications and challenges. arXiv preprint arXiv:2407.09618

2024 arXiv
[40]

Pan, Q.; Zhu, Y.; and Chu, L. 2023. Lumos: Heterogeneity-aware federated graph learning over decentralized devices. In 2023 IEEE 39th International Conference on Data Engineering (ICDE), 1914--1926. IEEE

2023
[41]

Qian, J.; Wan, G.; Huang, W.; Zhang, G.; Wu, Y.; Du, B.; and Ye, M. 2025. GHOST: Generalizable One-Shot Federated Graph Learning with Proxy-Based Topology Knowledge Retention. In Forty-second International Conference on Machine Learning

2025
[42]

Rombach, R.; Blattmann, A.; Lorenz, D.; Esser, P.; and Ommer, B. 2022. High-resolution image synthesis with latent diffusion models. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 10684--10695

2022
[43]

Ross, T.-Y.; and Doll \'a r, G. 2017. Focal loss for dense object detection. In proceedings of the IEEE conference on computer vision and pattern recognition, 2980--2988

2017
[44]

Shchur, O.; Mumme, M.; Bojchevski, A.; and G \"u nnemann, S. 2018. Pitfalls of graph neural network evaluation. arXiv preprint arXiv:1811.05868

2018 arXiv
[45]

Z.; Festag, A.; Trinitis, C.; Schulz, M.; and Knoll, A

Song, R.; Liu, D.; Chen, D. Z.; Festag, A.; Trinitis, C.; Schulz, M.; and Knoll, A. 2023. Federated learning via decentralized dataset distillation in resource-constrained edge environments. In 2023 International Joint Conference on Neural Networks (IJCNN), 1--10. IEEE

2023
[46]

Tan, Y.; Liu, Y.; Long, G.; Jiang, J.; Lu, Q.; and Zhang, C. 2023. Federated learning on non-iid graphs via structural knowledge sharing. In Proceedings of the AAAI conference on artificial intelligence, volume 37, 9953--9961

2023
[47]

Tan, Z.; Wan, G.; Huang, W.; and Ye, M. 2024. Fedssp: Federated graph learning with spectral knowledge and personalized preference. Advances in Neural Information Processing Systems, 37: 34561--34581

2024
[48]

Tang, T.; Han, Z.; Cai, Z.; Yu, S.; Zhou, X.; Oseni, T.; and Das, S. K. 2024. Personalized Federated Graph Learning on Non-IID Electronic Health Records. IEEE Transactions on Neural Networks and Learning Systems

2024
[49]

Veli c kovi \'c , P.; Cucurull, G.; Casanova, A.; Romero, A.; Lio, P.; and Bengio, Y. 2017. Graph attention networks. arXiv preprint arXiv:1710.10903

2017 arXiv
[50]

Wang, B.; Li, A.; Pang, M.; Li, H.; and Chen, Y. 2022. Graphfl: A federated learning framework for semi-supervised node classification on graphs. In 2022 IEEE International Conference on Data Mining (ICDM), 498--507. IEEE

2022
[51]

Wang, D.; Li, C.; Wen, S.; Nepal, S.; and Xiang, Y. 2020. Man-in-the-middle attacks against machine learning classifiers via malicious generative models. IEEE Transactions on Dependable and Secure Computing, 18(5): 2074--2087

2020
[52]

Wu, C.; Wu, F.; Cao, Y.; Huang, Y.; and Xie, X. 2021. Fedgnn: Federated graph neural network for privacy-preserving recommendation. arXiv preprint arXiv:2102.04925

2021 arXiv
[53]

Wu, F.; Souza, A.; Zhang, T.; Fifty, C.; Yu, T.; and Weinberger, K. 2019. Simplifying graph convolutional networks. In International conference on machine learning, 6861--6871. PMLR

2019
[54]

Wu, Z.; Pan, S.; Chen, F.; Long, G.; Zhang, C.; and Philip, S. Y. 2020. A comprehensive survey on graph neural networks. IEEE transactions on neural networks and learning systems, 32(1): 4--24

2020
[55]

Xiao, Z.; Wang, Y.; Liu, S.; Wang, H.; Song, M.; and Zheng, T. 2024. Simple graph condensation. In Joint European Conference on Machine Learning and Knowledge Discovery in Databases, 53--71. Springer

2024
[56]

Xie, H.; Ma, J.; Xiong, L.; and Yang, C. 2021. Federated graph classification over non-iid graphs. Advances in neural information processing systems, 34: 18839--18852

2021
[57]

Yan, B.; Cao, Y.; Wang, H.; Yang, W.; Du, J.; and Shi, C. 2024. Federated heterogeneous graph neural network for privacy-preserving recommendation. In Proceedings of the ACM on Web Conference 2024, 3919--3929

2024
[58]

Yang, M.; Su, S.; Li, B.; and Xue, X. 2024 a . Exploring One-Shot Semi-supervised Federated Learning with Pre-trained Diffusion Models. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, 16325--16333

2024
[59]

Yang, M.; Su, S.; Li, B.; and Xue, X. 2024 b . FedDEO: Description-Enhanced One-Shot Federated Learning with Diffusion Models. arXiv preprint arXiv:2407.19953

2024 arXiv
[60]

Yang, Z.; Cohen, W.; and Salakhudinov, R. 2016. Revisiting semi-supervised learning with graph embeddings. In International conference on machine learning, 40--48. PMLR

2016
[61]

Yao, Y.; Jin, W.; Ravi, S.; and Joe-Wong, C. 2024. FedGCN: Convergence-communication tradeoffs in federated training of graph convolutional networks. Advances in neural information processing systems, 36

2024
[62]

M.; Li, Z.; Mallya, A.; Hoiem, D.; Jha, N

Yin, H.; Molchanov, P.; Alvarez, J. M.; Li, Z.; Mallya, A.; Hoiem, D.; Jha, N. K.; and Kautz, J. 2020. Dreaming to distill: Data-free knowledge transfer via deepinversion. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 8715--8724

2020
[63]

Zeng, H.; Xu, M.; Zhou, T.; Wu, X.; Kang, J.; Cai, Z.; and Niyato, D. 2024. One-shot-but-not-degraded Federated Learning. In Proceedings of the 32nd ACM International Conference on Multimedia, 11070--11079

2024
[64]

Zeng, H.; Zhou, H.; Srivastava, A.; Kannan, R.; and Prasanna, V. 2019. Graphsaint: Graph sampling based inductive learning method. arXiv preprint arXiv:1907.04931

2019 arXiv
[65]

Zhang, C.; Long, G.; Zhou, T.; Zhang, Z.; Yan, P.; and Yang, B. 2024 a . GPFedRec: Graph-Guided Personalization for Federated Recommendation. In Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 4131--4142

2024
[66]

Zhang, H.; Li, X.; Zhu, Y.; and Hu, L. 2025. Rethinking Federated Graph Learning: A Data Condensation Perspective. arXiv preprint arXiv:2505.02573

2025 arXiv
[67]

Zhang, J.; Chen, C.; Li, B.; Lyu, L.; Wu, S.; Ding, S.; Shen, C.; and Wu, C. 2022. Dense: Data-free one-shot federated learning. Advances in Neural Information Processing Systems, 35: 21414--21428

2022
[68]

M.; and Yang, C

Zhang, K.; Sun, L.; Ding, B.; Yiu, S. M.; and Yang, C. 2024 b . Deep Efficient Private Neighbor Generation for Subgraph Federated Learning. In Proceedings of the 2024 SIAM International Conference on Data Mining (SDM), 806--814. SIAM

2024
[69]

Zhang, K.; Yang, C.; Li, X.; Sun, L.; and Yiu, S. M. 2021. Subgraph federated learning with missing neighbor generation. Advances in Neural Information Processing Systems, 34: 6671--6682

2021
[70]

Zhao, Y.; Li, M.; Lai, L.; Suda, N.; Civin, D.; and Chandra, V. 2018. Federated learning with non-iid data. arXiv preprint arXiv:1806.00582

2018 arXiv
[71]

Zhou, Y.; Pu, G.; Ma, X.; Li, X.; and Wu, D. 2020. Distilled one-shot federated learning. arXiv preprint arXiv:2009.07999

2020 arXiv
[72]

Zhu, G.; Li, D.; Gu, H.; Han, Y.; Yao, Y.; Fan, L.; and Yang, Q. 2024 a . Evaluating Membership Inference Attacks and Defenses in Federated Learning. arXiv preprint arXiv:2402.06289

2024 arXiv
[73]

Zhu, H.; and Koniusz, P. 2021. Simple spectral graph convolution. In International conference on learning representations

2021
[74]

Zhu, Y.; Li, X.; Wu, Z.; Wu, D.; Hu, M.; and Li, R.-H. 2024 b . FedTAD: Topology-aware Data-free Knowledge Distillation for Subgraph Federated Learning. arXiv preprint arXiv:2404.14061

2024 arXiv
[75]

Zhu, Z.; Hong, J.; and Zhou, J. 2021. Data-free knowledge distillation for heterogeneous federated learning. In International conference on machine learning, 12878--12889. PMLR

2021

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.

[1] [1]

, " * write output.state after.block = add.period write newline

ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.a...

[2] [2]

write newline

" write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...

[3] [3]

Abu-El-Haija, S.; Perozzi, B.; Kapoor, A.; Alipourfard, N.; Lerman, K.; Harutyunyan, H.; Ver Steeg, G.; and Galstyan, A. 2019. Mixhop: Higher-order graph convolutional architectures via sparsified neighborhood mixing. In international conference on machine learning, 21--29. PMLR

work page 2019

[4] [4]

S.; and Conti, M

Acar, A.; Aksu, H.; Uluagac, A. S.; and Conti, M. 2018. A survey on homomorphic encryption schemes: Theory and implementation. ACM Computing Surveys (Csur), 51(4): 1--35

work page 2018

[5] [5]

Aliakbari, J.; \"O stman, J.; et al. 2024. Decoupled Subgraph Federated Learning. arXiv preprint arXiv:2402.19163

work page Pith review arXiv 2024

[6] [6]

Baek, J.; Jeong, W.; Jin, J.; Yoon, J.; and Hwang, S. J. 2023. Personalized subgraph federated learning. In International conference on machine learning, 1396--1415. PMLR

2023

[7] [7]

Bang, D.; Lim, S.; Lee, S.; and Kim, S. 2023. Biomedical knowledge graph learning for drug repurposing by extending guilt-by-association to multiple layers. Nature Communications, 14(1): 3570

2023

[8] [8]

B.; Patel, S.; Ramage, D.; Segal, A.; and Seth, K

Bonawitz, K.; Ivanov, V.; Kreuter, B.; Marcedone, A.; McMahan, H. B.; Patel, S.; Ramage, D.; Segal, A.; and Seth, K. 2017. Practical secure aggregation for privacy-preserving machine learning. In proceedings of the 2017 ACM SIGSAC Conference on Computer and Communications Security, 1175--1191

2017