Pith. sign in

REVIEW 4 major objections 4 minor 47 references

Streamlined Federated Unlearning: Unite as One to Be Highly Efficient

T0 review · 4 major / 4 minor · reviewed 2026-08-12 · deepseek-v4-flash

Pith's one-line read SFU forgets target classes in 1-2 rounds, matching retraining

desk verdict A plausible multi-teacher unlearning idea undermined by incorrect theory and results that contradict its own mechanism. read the letter →

arxiv 2412.00126 v2 pith:MTAQA2MW submitted 2024-11-28 cs.LG

classification cs.LG
keywords federatedunlearningmulti-teacherdistillationknowledgerighttobeforgottenlearningbackdoorattackclass-levelcommunicationefficiency
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 claims that federated unlearning can be compressed into a single step: instead of first erasing a class's influence and then fine-tuning to recover accuracy, one distillation objective does both at once. The proposed SFU method gives each client three teacher models — a copy of the original model, a randomly initialized copy, and one-hot labels for retained data — and trains the unlearned model to match them on forgotten and retained data respectively. The authors report that on CIFAR-10, CIFAR-100, and DBpedia, this removes target-class influence to roughly the retraining baseline while keeping retained-data accuracy stable, in one or two communication rounds with 40x-152x time speedups over retraining. The reason a sympathetic reader should care is that, if true, the unlearning objective itself replaces the two-phase procedure that dominates current federated unlearning methods.

What carries the argument

The load-bearing object is the multi-teacher distillation system and its objective, Eq. (12). Three teachers are involved: $M^P$, an identical copy of the original model that anchors performance on retained data; $M^F$, a randomly initialized copy of the architecture that stands for a model which never saw the forgotten data; and $M^L_i$, one-hot encodings of the true labels of each client's retained data. The student model starts as the original model and is updated by minimizing KL divergence toward these teachers, so that the forgotten data are pulled toward the random teacher's behavior while retained data are pulled toward the original and its labels. The essential identity is the claim that minimizing divergence to $M^F$ drives the student's predictions on the forgotten class to uniformity, which is what converts an unlearning objective into actual influence removal.

What would settle it

Measure the softmax outputs of a freshly randomly initialized copy of the trained model on the to-be-forgotten class; if the mean maximum probability is clearly above $1/K$ (or the distribution is far from uniform), the premise behind Eq. (14) fails, and a model matched to that teacher would not be equivalent to a retrained model.

Watch

Extended reading notes

Core claim

SFU's central claim is that the multi-teacher loss in Eq. (12) achieves both goals of federated unlearning in a single step. On each client the student model is initialized to the original model and then minimizes two KL terms whose teacher signals are sums of teacher distributions: $\mathrm{KL}\big(M^U_i(x_i)\,\big\|\,\alpha M^F(x^f_i)+M^P(x^r_i)\big)$ plus $\mathrm{KL}\big(M^U_i(x_i)\,\big\|\,\alpha M^F(x^f_i)+M^L_i(x^r_i)\big)$. The forgetting teacher $M^F$ is a randomly initialized copy of the architecture; the paper argues that because it has never seen the forgotten data $x^f_i$, its output is approximately uniform, so driving the student's KL to zero makes the unlearned model's predictions on the forgotten class uniform and thereby erases that data's influence. The preservation teachers $M^P$ (a copy of the original model) and $M^L_i$ (one-hot labels of the retained data) hold the student's behavior on retained data steady. On CIFAR-10, CIFAR-100, and DBpedia the result is forgotten-data accuracy and backdoor success rates close to retraining, retained-data accuracy within a point or so of the original model, and completion in one or two communication rounds.

Load-bearing premise

The load-bearing premise is that a randomly initialized copy of the model produces near-uniform predictions on data it never trained on, so matching that teacher's outputs erases the target data's influence; if initialized networks do not behave this way, the theoretical argument collapses even if forgotten-class accuracy happens to fall to zero.

Editorial extensions

If this is right

  • Class-level federated unlearning can be completed in one or two communication rounds without a separate fine-tuning phase.
  • Retained-data accuracy stays close to the pre-unlearning model while forgotten-data accuracy approaches the retraining baseline, so forgetting does not visibly trade off against utility.
  • Backdoor attack success rates on forgotten data drop to roughly the retraining level, indicating that implanted behaviors on the target class are removed.
  • The same update works for single-class and multi-class requests and transfers from convolutional image models to an LSTM text model.
  • No historical update logs are needed, sidestepping the storage cost of many existing federated unlearning designs.

Reading between the lines

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

  • Editorial extension: the random-teacher-as-forgetter recipe is a distillation analogue of gradient ascent; if random initialization is not uniform enough, calibrating $M^F$'s outputs or substituting a model trained on disjoint proxy data would be a natural variant to test.
  • Editorial extension: the paper verifies forgetting with classification accuracy and backdoor success rate, but not with representation-level or membership-inference tests; those could reveal residual influence even when the uniform-output condition is met.
  • Editorial extension: because Eq. (12) feeds unnormalized sums of teacher distributions into KL divergences, a mixture-of-teachers formulation with weights summing to one is a direct test of whether the reported balance between forgetting and preservation depends on that normalization choice.
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. The paper proposes SFU, a federated unlearning method that unifies influence removal and performance preservation in a single multi-teacher distillation step. Three teacher models (a copy of the original model M^P, a randomly initialized model M^F, and a one-hot label teacher M^L_i) guide the unlearned student model on each client; the server aggregates the resulting models. The authors report that SFU removes target-class influence as effectively as retraining while preserving retained-data accuracy, in only 1-2 communication rounds, with time speedups of 40-152x over retraining on CIFAR-10, CIFAR-100, and DBpedia. A theoretical analysis in Section V claims that minimizing a KL-based loss drives the student's predictions on forgotten data to a uniform distribution, thereby erasing the target class.

Significance. If the central claim held, SFU would be a practically appealing contribution: it is simple, avoids historical-information storage, works without global data access, and combines unlearning and performance recovery in one step. The experimental breadth (three datasets, multiple target classes, backdoor verification) and the very large reported efficiency gains are concrete strengths. However, the significance is currently undermined by a theoretical analysis that contains incorrect gradient formulas, an unverified and internally inconsistent assumption about random teacher outputs, and a lack of controls isolating the multi-teacher loss from the subsequent data deletion. The claim that the multi-teacher system itself, rather than data removal and resumed FL training, produces the observed forgetting is not established.

major comments (4)
  1. [Section V.A, Eqs. (15), (17), (19)] The gradient formulas for KL divergence are mathematically incorrect. For KL(p||q) with q fixed, the derivative with respect to the student parameters θ is Σ_k (∂p_k/∂θ)(log(p_k/q_k)+1), not Σ_k (∂p_k/∂θ)(1−p_k/q_k). This error appears in all three gradient derivations and invalidates the stated convergence argument. The authors should either correct the derivations or remove the analytic claims and rely on empirical evaluation.
  2. [Section V.A, Eq. (14) vs. Table III] The mechanism described in Eq. (14) is inconsistent with the reported results. If minimizing D_forget drives M^U_i(x^f_i) to Uniform(K), then top-1 accuracy on forgotten data should approach 1/K: about 10% on CIFAR-10, 1% on CIFAR-100, and 7% on DBpedia. Table III reports SFU forgotten-data accuracies of 0.11-0.67%, 0.12-0.47%, and 0.04-0.24%, respectively, which are one to two orders of magnitude below chance. Matching a randomly initialized teacher (whose softmax outputs are near-uniform) cannot produce below-chance accuracy. The paper provides no measurement of M^F's actual output distribution on forgotten samples, so the stated theoretical mechanism cannot explain the headline numbers.
  3. [Section IV.C, Eq. (12)] The combined teacher distributions in Eq. (12), e.g., M^P(x^r_i) + α·M^F(x^f_i), are not normalized probability distributions, and the notation mixes arguments over different data subsets. KL divergence requires valid probability distributions over the same support. The authors need to specify the exact per-sample loss, including how the terms over retained and forgotten data are combined and normalized, before the objective can be meaningfully analyzed.
  4. [Algorithm 1 and Section IV.D] In Algorithm 1, the forgotten data x^f_i are deleted only after the unlearning loop, and then normal FL training resumes. Since the reported efficiency is 1-2 communication rounds, the measured forgetting could be substantially caused by the deletion of x^f_i and the subsequent resumed FL training, rather than by the multi-teacher loss itself. A control experiment that isolates the multi-teacher loss (e.g., running the loss without deleting x^f_i, or deleting x^f_i without running the loss) is necessary to attribute the observed forgetting to the proposed objective. Without such a control, the central claim that the multi-teacher system achieves both goals is not demonstrated.
minor comments (4)
  1. [Section IV.B, Eq. (8)] Equation (8) states M^P = random initialize(M^O), but M^P is defined as a copy of the original model; this should be M^F = random initialize(M^O). The current equation contradicts the definition in the bulleted text.
  2. [Section IV.C, paragraph before Eq. (9)] The sentence defining x^r_i and x^f_i is reversed: it says x^r_i is 'the target data to be forgotten' and x^f_i is 'the non-target data to be retained', which is the opposite of the notation in Table II and the rest of the paper.
  3. [Table III and Figure 3] The tables and figures report point estimates without standard deviations or confidence intervals across multiple runs. Given the small differences between SFU and retraining on retained accuracy, some measure of variability is needed to support the claim of no degradation.
  4. [Throughout] There are several typographical issues: 'Dbsepia' in Table III and elsewhere, 'Unlearnng' in Section VI.B.3, and the term 'isomorphic' in Section IV.D should likely be 'homogeneous'.

Circularity Check

0 steps flagged · score 0.0 of 10

No material circularity: the multi-teacher loss is the algorithm's stated objective, and the reported gains are empirical evaluations rather than quantities derived from fitted inputs or self-citations.

full rationale

We examined the derivation chain for SFU and found no step in which a claimed prediction is equivalent to its input by construction. The three teachers in Eq. (12) are explicit algorithmic inputs, not parameters fitted to the unlearning outcomes: M^P is the original model by definition, M^F is a randomly initialized copy, M^L_i is the one-hot label encoding of retained data, and alpha is set by the data-size ratio rather than by fitting to unlearning results. The preserved retained-data accuracy is therefore a direct consequence of the loss being minimized, since the objective contains KL terms pulling the student toward M^P and M^L on retained data; reporting that retention is achieved is an evaluation of the stated design, not a fitted quantity relabeled as a prediction. The unlearning-side analysis in Section V.A is informal and contains an unverified assumption that M^F's outputs are approximately uniform, and the gradient expressions in Eqs. (15), (17), and (19) are mathematically incorrect for KL divergences, but these are soundness and correctness concerns rather than circularity because the theoretical argument does not assume the reported near-zero forgotten accuracies. In fact, the uniform-teacher story would predict chance-level rather than near-zero accuracies, so the experimental numbers are not manufactured by the equations. We also found no load-bearing self-citation: the references are to external prior work, and no uniqueness theorem from the authors is invoked to force the design. Thus the central claims rest on experiments and an explicitly stated objective, not on a derivation that reduces to its own inputs.

Assumptions & free parameters 2 free parameters · 4 assumptions · 1 invented entities

The central claim rests on treating a randomly initialized model as a uniform-distribution oracle for forgetting, and on treating unnormalized sums of teacher outputs as valid KL targets. These are untested assumptions that the paper asserts rather than derives. Alpha and T are the only hand-set numeric parameters, with alpha tied to the class ratio.

free parameters (2)
  • alpha = num(xr)/num(xf) = dataset-dependent (e.g., 9 for one-class unlearning in CIFAR-10)
    Balances the forgetting teacher against preservation teachers in Eq. (12); chosen from data ratio and validated by ablation, not learned.
  • Distillation temperature T = 1
    Set constant for all experiments; controls the softness of teacher outputs but is not tuned.
assumptions (4)
  • ad hoc to paper Randomly initialized deep networks output approximately uniform class distributions (Eq. 14).
    Underpins the claim that matching M^F erases target data; not generally true and unverified.
  • ad hoc to paper The combined teacher distribution (e.g., M^P + alpha M^F) is a valid probability distribution for KL divergence (Eq. 12).
    Sum of distributions is not normalized, so the loss as written is not a proper KL divergence.
  • domain assumption All clients share the same model architecture (isomorphic environment).
    Stated in Section IV.D; needed for server-side aggregation of locally unlearned models.
  • standard math FedAvg aggregation and standard SGD are sufficient for FL training and unlearning.
    Used throughout; standard in the field.
invented entities (1)
  • Forgetting teacher model M^F (randomly initialized copy of original architecture)
    purpose: Acts as a stand-in for a model that has never seen target data; the student is pushed toward M^F's outputs on forgotten samples.
    The premise that its outputs are approximately uniform is central, untested, and generally false for deep networks.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Streamlined Federated Unlearning: Unite as One to Be Highly Efficient." pith.science (2026). https://pith.science/paper/MTAQA2MW

@misc{pith2026241200126,
  author       = {Pith},
  title        = {Pith review of: Streamlined Federated Unlearning: Unite as One to Be Highly Efficient},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MTAQA2MW}},
  note         = {Machine review of arXiv:2412.00126}
}
read the original abstract

Recently, the enactment of ``right to be forgotten" laws and regulations has imposed new privacy requirements on federated learning (FL). Researchers aim to remove the influence of certain data from the trained model without training from scratch through federated unlearning (FU). While current FU research has shown progress in enhancing unlearning efficiency, it often results in degraded model performance upon achieving the goal of data unlearning, necessitating additional steps to recover the performance of the unlearned model. Moreover, these approaches also suffer from many shortcomings such as high consumption of computational and storage resources. To this end, we propose a streamlined federated unlearning approach (SFU) aimed at effectively removing the influence of the target data while preserving the model performance on the retained data without degradation. We design a practical multi-teacher system that achieves both target data influence removal and model performance preservation by guiding the unlearned model through several distinct teacher models. SFU is both computationally and storage-efficient, highly flexible, and generalizable. We conduct extensive experiments on both image and text benchmark datasets. The results demonstrate that SFU significantly improves time and communication efficiency compared to the benchmark retraining method and significantly outperforms existing SOTA methods. Additionally, we verify the effectiveness of SFU using the backdoor attack.

Figures

Figures reproduced from arXiv: 2412.00126 by the authors.

Figure 1
Figure 1. The general process of federated unlearning. [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. The framework of SFU. When an unlearning request reaches the server, each client (client [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Test accuracy of models using different unlearning methods on forgotten and retained data. Three different classes of [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: The unlearning performance of SFU under different combinations of teacher models. The lines of different colors [PITH_FULL_IMAGE:figures/full_fig_p011_4.png]
Figure 5
Figure 5. Figure 5: Comparison of the unlearning performance of SFU with different values of hyperparameters [PITH_FULL_IMAGE:figures/full_fig_p011_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

47 extracted references · 23 canonical work pages

  1. [1]

    Scu: An efficient machine unlearning scheme for deep learning enabled semantic communications,

    W. Wang, Z. Tian, C. Zhang, and S. Yu, “Scu: An efficient machine unlearning scheme for deep learning enabled semantic communications,” IEEE Transactions on Information Forensics and Security , vol. 20, pp. 547–558, 2025

  2. [2]

    Zero- shot machine unlearning,

    V . S. Chundawat, A. K. Tarun, M. Mandal, and M. Kankanhalli, “Zero- shot machine unlearning,” IEEE Transactions on Information Forensics and Security, vol. 18, pp. 2345–2354, 2023

  3. [3]

    Communication-efficient learning of deep networks from decentralized data,

    B. McMahan, E. Moore, D. Ramage, S. Hampson, and B. A. y Arcas, “Communication-efficient learning of deep networks from decentralized data,” in Artificial intelligence and statistics . PMLR, 2017, pp. 1273– 1282

  4. [4]

    General data protection regulation (gdpr),

    European Union, “General data protection regulation (gdpr),” Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 April 2016, 2016, available at https://eur-lex.europa.eu/legal-content/ EN/TXT/?uri=CELEX%3A32016R0679

  5. [5]

    Klein, Canadian Privacy: Data Protection and Policy for the Practi- tioner

    K. Klein, Canadian Privacy: Data Protection and Policy for the Practi- tioner. International Association of Privacy Professionals, 2020, focuses on data protection laws and policies in Canada

  6. [6]

    A survey on federated unlearning: Challenges, methods, and future directions,

    Z. Liu, Y . Jiang, J. Shen, M. Peng, K.-Y . Lam, X. Yuan, and X. Liu, “A survey on federated unlearning: Challenges, methods, and future directions,” ACM Computing Surveys , vol. 57, no. 1, pp. 1–38, 2024

  7. [7]

    Machine unlearning,

    L. Bourtoule, V . Chandrasekaran, C. A. Choquette-Choo, H. Jia, A. Travers, B. Zhang, D. Lie, and N. Papernot, “Machine unlearning,” in 2021 IEEE Symposium on Security and Privacy (SP) . IEEE, 2021, pp. 141–159

  8. [8]

    Can bad teaching induce forgetting? unlearning in deep networks using an incompetent teacher,

    V . S. Chundawat, A. K. Tarun, M. Mandal, and M. Kankanhalli, “Can bad teaching induce forgetting? unlearning in deep networks using an incompetent teacher,” in Proceedings of the AAAI Conference on Artificial Intelligence, vol. 37, no. 6, 2023, pp. 7210–7217

Show all 47 references
  1. [9]

    Machine unlearning via rep- resentation forgetting with parameter self-sharing,

    W. Wang, C. Zhang, Z. Tian, and S. Yu, “Machine unlearning via rep- resentation forgetting with parameter self-sharing,” IEEE Transactions on Information Forensics and Security , vol. 19, pp. 1099–1111, 2023

  2. [10]

    Forgetting and remembering are both you need: Balanced graph structure unlearning,

    C. Zhang, W. Wang, Z. Tian, and S. Yu, “Forgetting and remembering are both you need: Balanced graph structure unlearning,” IEEE Trans- actions on Information Forensics and Security , vol. 19, pp. 6751–6763, 2024

  3. [11]

    Sok: Challenges and opportu- nities in federated unlearning,

    H. Jeong, S. Ma, and A. Houmansadr, “Sok: Challenges and opportu- nities in federated unlearning,” arXiv preprint arXiv:2403.02437 , 2024

  4. [12]

    Fedrecover: Recovering from poisoning attacks in federated learning using historical information,

    X. Cao, J. Jia, Z. Zhang, and N. Z. Gong, “Fedrecover: Recovering from poisoning attacks in federated learning using historical information,” in 2023 IEEE Symposium on Security and Privacy (SP) . IEEE, 2023, pp. 1366–1383

  5. [13]

    Revfrf: Enabling cross-domain random forest training with revocable federated learning,

    Y . Liu, Z. Ma, Y . Yang, X. Liu, J. Ma, and K. Ren, “Revfrf: Enabling cross-domain random forest training with revocable federated learning,” IEEE Transactions on Dependable and Secure Computing, vol. 19, no. 6, pp. 3671–3685, 2021

  6. [14]

    Mit- igating poor data quality impact with federated unlearning for human- centric metaverse,

    P. Wang, Z. Wei, H. Qi, S. Wan, Y . Xiao, G. Sun, and Q. Zhang, “Mit- igating poor data quality impact with federated unlearning for human- centric metaverse,” IEEE Journal on Selected Areas in Communications, vol. 42, no. 4, pp. 832–849, 2024

  7. [15]

    Unlearning backdoor attacks in federated learning,

    C. Wu, S. Zhu, and P. Mitra, “Unlearning backdoor attacks in federated learning,” in ICLR 2023 Workshop on Backdoor Attacks and Defenses in Machine Learning , 2024, pp. 1–9

  8. [16]

    Fedrecov- ery: Differentially private machine unlearning for federated learning frameworks,

    L. Zhang, T. Zhu, H. Zhang, P. Xiong, and W. Zhou, “Fedrecov- ery: Differentially private machine unlearning for federated learning frameworks,” IEEE Transactions on Information Forensics and Security, vol. 18, pp. 4732–4746, 2023

  9. [17]

    Forgettable federated linear learning with certified data removal,

    R. Jin, M. Chen, Q. Zhang, and X. Li, “Forgettable federated linear learning with certified data removal,” arXiv preprint arXiv:2306.02216 , 2023

  10. [18]

    The right to be forgotten in federated learning: An efficient realization with rapid retraining,

    Y . Liu, L. Xu, X. Yuan, C. Wang, and B. Li, “The right to be forgotten in federated learning: An efficient realization with rapid retraining,” in IEEE INFOCOM 2022-IEEE Conference on Computer Communications. IEEE, 2022, pp. 1749–1758

  11. [19]

    Federated unlearning via active forgetting,

    Y . Li, C. Chen, X. Zheng, and J. Zhang, “Federated unlearning via active forgetting,” arXiv preprint arXiv:2307.03363 , 2023

  12. [20]

    Quickdrop: Efficient federated unlearning by integrated dataset distillation,

    A. Dhasade, Y . Ding, S. Guo, A.-m. Kermarrec, M. De V os, and L. Wu, “Quickdrop: Efficient federated unlearning by integrated dataset distillation,” arXiv preprint arXiv:2311.15603 , 2023

  13. [21]

    Forget-svgd: Particle- based bayesian federated unlearning,

    J. Gong, J. Kang, O. Simeone, and R. Kassab, “Forget-svgd: Particle- based bayesian federated unlearning,” in 2022 IEEE Data Science and Learning Workshop (DSLW). IEEE, 2022, pp. 1–6

  14. [22]

    Compressed particle-based feder- ated bayesian learning and unlearning,

    J. Gong, O. Simeone, and J. Kang, “Compressed particle-based feder- ated bayesian learning and unlearning,” IEEE Communications Letters , vol. 27, no. 2, pp. 556–560, 2022

  15. [23]

    Fedharmony: Unlearning scanner bias with distributed data,

    N. K. Dinsdale, M. Jenkinson, and A. I. Namburete, “Fedharmony: Unlearning scanner bias with distributed data,” in International Confer- ence on Medical Image Computing and Computer-Assisted Intervention . Springer, 2022, pp. 695–704

  16. [24]

    Bfu: Bayesian federated unlearning with parameter self-sharing,

    W. Wang, Z. Tian, C. Zhang, A. Liu, and S. Yu, “Bfu: Bayesian federated unlearning with parameter self-sharing,” in Proceedings of the 2023 ACM Asia Conference on Computer and Communications Security, 2023, pp. 567–578. 13

  17. [25]

    Heterogeneous federated knowledge graph embedding learning and unlearning,

    X. Zhu, G. Li, and W. Hu, “Heterogeneous federated knowledge graph embedding learning and unlearning,” in Proceedings of the ACM web conference 2023, 2023, pp. 2444–2454

  18. [26]

    Asynchronous federated unlearning,

    N. Su and B. Li, “Asynchronous federated unlearning,” in IEEE INFO- COM 2023-IEEE Conference on Computer Communications . IEEE, 2023, pp. 1–10

  19. [27]

    Federaser: Enabling efficient client-level data removal from federated learning models,

    G. Liu, X. Ma, Y . Yang, C. Wang, and J. Liu, “Federaser: Enabling efficient client-level data removal from federated learning models,” in 2021 IEEE/ACM 29th international symposium on quality of service (IWQOS). IEEE, 2021, pp. 1–10

  20. [28]

    Towards efficient and certified recovery from poisoning attacks in federated learning,

    Y . Jiang, J. Shen, Z. Liu, C. W. Tan, and K.-Y . Lam, “Towards efficient and certified recovery from poisoning attacks in federated learning,” arXiv preprint arXiv:2401.08216 , 2024

  21. [29]

    Com- munication efficient and provable federated unlearning,

    Y . Tao, C.-L. Wang, M. Pan, D. Yu, X. Cheng, and D. Wang, “Com- munication efficient and provable federated unlearning,” arXiv preprint arXiv:2401.11018, 2024

  22. [30]

    Federated unlearning for on-device recommendation,

    W. Yuan, H. Yin, F. Wu, S. Zhang, T. He, and H. Wang, “Federated unlearning for on-device recommendation,” in Proceedings of the six- teenth ACM international conference on web search and data mining , 2023, pp. 393–401

  23. [31]

    Fast federated machine unlearning with nonlinear functional theory,

    T. Che, Y . Zhou, Z. Zhang, L. Lyu, J. Liu, D. Yan, D. Dou, and J. Huan, “Fast federated machine unlearning with nonlinear functional theory,” in International conference on machine learning. PMLR, 2023, pp. 4241– 4268

  24. [32]

    A revocation key- based approach towards efficient federated unlearning,

    R.-Z. Xu, S.-Y . Hong, P.-W. Chi, and M.-H. Wang, “A revocation key- based approach towards efficient federated unlearning,” in 2023 18th Asia Joint Conference on Information Security (AsiaJCIS). IEEE, 2023, pp. 17–24

  25. [33]

    Edge caching with federated unlearning for low-latency v2x communications,

    P. Wang, Z. Yan, M. S. Obaidat, Z. Yuan, L. Yang, J. Zhang, Z. Wei, and Q. Zhang, “Edge caching with federated unlearning for low-latency v2x communications,” IEEE Communications Magazine, vol. 62, no. 10, pp. 118–124, 2024

  26. [34]

    Verifi: Towards verifiable federated unlearning,

    X. Gao, X. Ma, J. Wang, Y . Sun, B. Li, S. Ji, P. Cheng, and J. Chen, “Verifi: Towards verifiable federated unlearning,” IEEE Transactions on Dependable and Secure Computing, vol. 21, no. 6, pp. 5720–5736, 2024

  27. [35]

    Federated unlearning with momentum degradation,

    Y . Zhao, P. Wang, H. Qi, J. Huang, Z. Wei, and Q. Zhang, “Federated unlearning with momentum degradation,” IEEE Internet of Things Journal, vol. 11, no. 5, pp. 8860–8870, 2024

  28. [36]

    Federated unlearning via class- discriminative pruning,

    J. Wang, S. Guo, X. Xie, and H. Qi, “Federated unlearning via class- discriminative pruning,” in Proceedings of the ACM Web Conference 2022, 2022, pp. 622–632

  29. [37]

    Federated un- learning: How to efficiently erase a client in fl?

    A. Halimi, S. Kadhe, A. Rawat, and N. Baracaldo, “Federated un- learning: How to efficiently erase a client in fl?” arXiv preprint arXiv:2207.05521, 2022

  30. [38]

    Subspace based federated unlearning,

    G. Li, L. Shen, Y . Sun, Y . Hu, H. Hu, and D. Tao, “Subspace based federated unlearning,” arXiv preprint arXiv:2302.12448 , 2023

  31. [39]

    Federated unlearning: Guarantee the right of clients to forget,

    L. Wu, S. Guo, J. Wang, Z. Hong, J. Zhang, and Y . Ding, “Federated unlearning: Guarantee the right of clients to forget,” IEEE Network , vol. 36, no. 5, pp. 129–135, 2022

  32. [40]

    Fedu: Federated unlearning via user-side influence approximation forgetting,

    W. Wang, C. Zhang, Z. Tian, and S. Yu, “Fedu: Federated unlearning via user-side influence approximation forgetting,” IEEE Transactions on Dependable and Secure Computing , pp. 1–14, 2024

  33. [41]

    Spfl: A self-purified federated learning method against poisoning attacks,

    Z. Liu, W. He, C.-H. Chang, J. Ye, H. Li, and X. Li, “Spfl: A self-purified federated learning method against poisoning attacks,” IEEE Transactions on Information Forensics and Security , vol. 19, pp. 6604–6619, 2024

  34. [42]

    Personal- ized edge intelligence via federated self-knowledge distillation,

    H. Jin, D. Bai, D. Yao, Y . Dai, L. Gu, C. Yu, and L. Sun, “Personal- ized edge intelligence via federated self-knowledge distillation,” IEEE Transactions on Parallel and Distributed Systems , vol. 34, no. 2, pp. 567–580, 2022

  35. [43]

    Learning multiple layers of features from tiny images,

    A. Krizhevsky, G. Hinton et al. , “Learning multiple layers of features from tiny images,” 2009

  36. [44]

    Character-level convolutional net- works for text classification,

    X. Zhang, J. Zhao, and Y . LeCun, “Character-level convolutional net- works for text classification,” Advances in neural information processing systems, vol. 28, 2015

  37. [45]

    Deep residual learning for image recognition,

    K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” in Proceedings of the IEEE conference on computer vision and pattern recognition , 2016, pp. 770–778

  38. [46]

    Long short-term memory,

    S. Hochreiter and J. Schmidhuber, “Long short-term memory,” Neural computation, vol. 9, no. 8, pp. 1735–1780, 1997

  39. [47]

    Bayesian nonparametric federated learning of neural networks,

    M. Yurochkin, M. Agarwal, S. Ghosh, K. Greenewald, N. Hoang, and Y . Khazaeni, “Bayesian nonparametric federated learning of neural networks,” in International conference on machine learning . PMLR, 2019, pp. 7252–7261

Pith tools

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