Pith. sign in

REVIEW 5 major objections 5 minor 55 references

Incentivizing High-quality Participation From Federated Learning Agents

T0 review · 5 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read The paper argues that a Wasserstein-distance measure of how non-IID each agent's data are can drive payments making truthful, high-effort participation a stable equilibrium in federated learning.

desk verdict The convergence-bound formalism is a real contribution, but the payment rule as written is not computable from server-visible information, and the equilibrium proof has a hole. read the letter →

arxiv 2506.16731 v1 pith:2QMZ3YNW submitted 2025-06-20 cs.AI cs.DCcs.LG

Jinlong Pang , Jiaheng Wei , Yifan Hua , Chen Qian , Yang Liu This is my paper · ORCID
classification cs.AIcs.DCcs.LG
keywords federatedlearningincentivemechanismdataheterogeneitynon-IIDdegreeWassersteindistancepeerpredictionStackelberggameNashequilibrium
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 show that a federated learning platform can pay each participating agent according to how far its local data are from a reference IID distribution, and that this payment rule makes truthful, high-effort participation a stable outcome. The non-IID degree of an agent is measured by the quantity the paper calls the Wasserstein distance, $\delta_k=\frac{1}{2}\sum_i |p^{(k)}(y=i)-p^{(c)}(y=i)|$, and the paper proves this quantity appears in the FL convergence bound and in the generalization-loss gap between any two agents. Rewards are built from a peer-prediction scoring function that pays an agent more when its model's loss gap to a randomly chosen peer is smaller. The interaction is formalized as a two-stage Stackelberg game, and the paper proves that a pure Nash equilibrium in effort levels exists when utility functions are well-behaved. If correct, this gives a route to rewarding data quality in FL without moving raw data off the agents.

What carries the argument

The load-bearing object is the quantity the paper calls the Wasserstein distance, defined by the label-proportion sum $\delta_k = \frac{1}{2}\sum_{i=1}^I |p^{(k)}(y=i)-p^{(c)}(y=i)|$. It appears in Theorem 2's divergence bound, so Example 1 becomes a FedAvg convergence bound with $\sum_k p_k\delta_k^2$ as an explicit cost of heterogeneity. The same $\delta_k$ enters Theorem 3's bound on the generalization-loss gap, $\Phi\delta_k^2+\Phi\delta_{k'}^2+\Upsilon$, and the payment rule inverts that bound: $f(Q/(\Phi\delta_k^2+\Phi\delta_{k'}^2+\Upsilon))$ pays more when the gap is predicted to be small. The randomly chosen peer $k'$, drawn from the peer-prediction approach, prevents coordinated strategic behavior, while the Stackelberg stage sets the payment coefficient $Q$ to maximize the learner's payoff.

What would settle it

Run the protocol with a server that has access only to model updates and a public reference distribution, and ask it to compute the payments; if the resulting payments do not match $f(Q/(\Phi\delta_k^2+\Phi\delta_{k'}^2+\Upsilon))$ for ground-truth $\delta$ values in a simulated setting, the mechanism is not implementable as specified.

Watch

Extended reading notes

Core claim

On its own terms, the paper's central discovery is that the non-IID degree $\delta_k$ is not just a description of heterogeneity but a quantity that controls both sides of the incentive problem. Theorem 2 bounds the divergence between local and averaged models by a constant multiple of $\sum_k p_k \delta_k^2$, which turns generic FL convergence bounds into bounds that explicitly penalize heterogeneous data; Theorem 3 bounds the generalization-loss gap between two agents by $\Phi\delta_k^2+\Phi\delta_{k'}^2+\Upsilon$. That gap bound is used to define the payment $Payment_k(e_k,e_{k'}) = f(Q/(\Phi\delta_k^2(e_k)+\Phi\delta_{k'}^2(e_{k'})+\Upsilon))$, so an agent earns more by reducing its own $\delta_k$ relative to its peer. The paper then models agents and learner as a Stackelberg game, characterizes each agent's optimal effort by a first-order condition, and proves that if utility functions are well-behaved, a pure Nash equilibrium in effort levels exists in which no agent can profitably deviate.

Load-bearing premise

The load-bearing premise is that the learner can actually evaluate the payment function, which requires knowing each agent's private label distribution or, equivalently, its non-IID degree $\delta_k$; the paper does not describe how the server obtains these values from the model updates it receives, so if they are unavailable the payment scheme cannot be computed.

Editorial extensions

If this is right

  • Theorem 2 converts the divergence term in existing FL convergence bounds into a weighted sum of squared non-IID degrees, so any agent that lowers $\delta_k$ directly tightens the global convergence bound.
  • The peer-prediction scoring rule ties payment to the generalization-loss gap with a randomly chosen peer, which the paper argues prevents coordinated untruthful behavior and makes high-effort participation stable.
  • The Stackelberg formulation gives the learner a principled way to set the initial payment coefficient $Q$ by maximizing its own payoff subject to non-negative agent utilities.
  • Because the scoring uses model responses and a public reference distribution, the mechanism is designed to reward data quality without requiring agents to reveal raw data.

Reading between the lines

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

  • The paper does not provide an inference procedure for $\delta_k$ from model updates; the most direct extension is to estimate $\delta_k$ from the server's observed gradient norms or predicted label distributions, and the mechanism's deployability depends on that estimate being accurate.
  • Because the payment is based on an upper bound rather than realized contribution, an agent could in principle game the bound by overfitting the reference validation set without improving its true non-IID degree; comparing payments to leave-one-out influence would test whether this matters.
  • The same $\delta_k$-to-convergence link could be used outside incentives, for example as a client-selection or data-valuation criterion in federated learning systems that already assume honest participation.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

5 major / 5 minor

Summary. The paper proposes an incentive mechanism for federated learning in which payments reward agents for reducing the distance between their local label distribution and a reference distribution. The authors reformulate the convergence bound for FedAvg using a total-variation-based non-IID degree (Theorem 2), derive an upper bound on the generalization-loss gap between any two agents (Theorem 3), and use that bound to define a payment function in Section 5.1. They model the interaction as a two-stage Stackelberg game and claim existence of a pure Nash equilibrium (Theorem 5). Experiments on MNIST, FashionMNIST, CIFAR-10, and CIFAR-100 illustrate the effect of non-IID degree on accuracy and show utility curves that are claimed to confirm the theoretical equilibrium.

Significance. If the mechanism were implementable, it would address a real gap in the FL incentive literature, where most schemes reward sample quantity rather than data quality or heterogeneity. The convergence-bound reformulation in Theorem 2 is a useful conceptual link, and the idea of using peer prediction to score generalization gaps is creative. The paper also provides fairly detailed proof sketches. However, the central payment rule is not computable from information the server actually holds, and the equilibrium proof contains a derivative error. These are not presentation issues; they block the main claims. As it stands, the contribution is a theoretical skeleton whose mechanism cannot be instantiated or tested, and the experimental validation is circular because the payment formula is fed the same private δ values it would need to observe.

major comments (5)
  1. [Section 5.1, payment function after Eq. (2)] The payment function is defined as Payment_k(e_k,e'_k) = f(Q/(Φδ_k^2(e_k)+Φδ_{k'}^2(e_{k'})+Υ)), where δ_k is the total-variation distance between agent k's private label distribution p^{(k)} and the reference p^{(c)} defined in Eq. (1). The learner only receives uploaded models and an auxiliary set D_c (Section 4.2); it never observes p^{(k)}. The statement that 'the learner can evaluate the value of δ_k' is unsupported, and no estimator from model updates or report-and-verify protocol is given. Consequently, the payment cannot be computed, agents cannot anticipate a well-defined utility, and the two-stage game in Section 5.2 is not a well-defined mechanism as stated. The experiments in Section 6 do not repair this: they construct clients with known δ values and feed those values into the same payment formula, so the utility curves in Figure 2 are true by construction rather than evidence that the mechanism is implementable.
  2. [Appendix A.10, Lemma 9] Lemma 9 claims that ∂f(e_k,e_{k'})/∂e_{k'} is the same as ∂f(e_k,e_{k'})/∂e_k. For the payment f(Q/(Φδ_k^2(e_k)+Φδ_{k'}^2(e_{k'})+Υ)), the derivative with respect to e_k depends on δ_k and δ'_k through Φδ_k^2, while the derivative with respect to e_{k'} depends on δ_{k'} and δ'_{k'}; these are not equal unless the two agents have identical effort-distance functions and effort levels. The bounds d1_{k'} and d2_k asserted in the proof therefore do not follow. Since the 'well-behaved' condition (Definition 3) is what Theorem 5 relies on, the existence proof for a pure Nash equilibrium is incomplete.
  3. [Theorem 4 and its proof (Section 5.2, Appendix A.9)] Theorem 4's statement says the optimal effort level e*_k is 1, but the displayed definition gives e*_k = 0 or Ŷe_k, not 1. In addition, the proof's first-order condition, ∂f(Ŷe_k,e_{k'})/∂δ_k + c d'(δ_k(Ŷe_k)) = 0, omits the chain-rule factor dδ_k/de_k and the sign of the derivative of the cost term c·d(|δ_k(0)-δ_k(e_k)|) that appears in Eq. (11). As written, the condition does not follow from the utility derivative. This affects the best-response characterization used in the backward-induction argument.
  4. [Theorem 2 versus Example 1] Theorem 2 states the bound as 16(E−1)G^2η_t^2(1+2η_tL)^{2(E−1)}Ψ with Ψ = ∑_k p_k (∑_i |p^{(k)}(y=i)−p^{(c)}(y=i)|)^2 = 4δ_k^2, which equals 64(E−1)G^2η_t^2(1+2η_tL)^{2(E−1)}∑_k p_k δ_k^2. Example 1 instead writes B = 16(E−1)G^2η_t^2(1+2η_tL)^{2(E−1)}∑_k p_k δ_k^2 and labels it 'Upper bound shown in Theorem 2'. These two expressions differ by a factor of 4, so the simplified bound used in the learner's payoff function is not justified by Theorem 2 as stated.
  5. [Assumptions in Theorem 2 and Lemma 6] Both Theorem 2 and Lemma 6 rely on the assumption that the per-class Lipschitz constants are equal, L = L_{x|y=i} for all i and i', but this assumption only appears inside the proofs in Appendix A.2 and A.4 and is not listed among Assumptions 1–4 in Section 3.1. If per-class Lipschitz constants differ, the displayed bounds do not follow. The assumption should be promoted to a numbered assumption in the main text, with a discussion of which loss functions satisfy it.
minor comments (5)
  1. [Eq. (1) and throughout] The quantity in Eq. (1) is the total variation distance (half the L1 distance between label distributions), not the Wasserstein distance. The terminology should be corrected throughout, or the Wasserstein formulation should actually be used and the proof adjusted accordingly.
  2. [Section 5.1] The notation is inconsistent: the payment is defined as Payment_k(e_k,e'_k) but then written as f(e_k,e_{k'}) in the text and in Assumption 5. Please use a single notation consistently.
  3. [Section 6 and Appendix B.4] The derivation of Υ in Appendix B.4 yields Υ = 2G^2/μ only when E is even and a negative value when E is odd. Since the paper later requires Υ > 0, the assumption that E is even should be stated in the main text where Theorem 3 is introduced, not only in the appendix.
  4. [Remark 2 / Remark 5] The claim that free-riding does not exist in the proposed setting is not established: the proof argues that the coefficient Q decouples the model's performance from agents' rewards, but the argument implicitly assumes uniqueness of the Nash equilibrium, which is not proven.
  5. [Section 6.2, Figure 2] The claim that utility 'remains stable' as training progresses is not evidence of a Nash equilibrium, because the payment formula depends only on the known δ values and the randomly selected peer, not on the training dynamics; the stability is essentially built into the formula. The experiments should test an actual strategic interaction or compare against a baseline mechanism.

Circularity Check

2 steps flagged · score 6.0 of 10

The payment rule is defined directly on the private non-IID degree δ_k that the mechanism is supposed to elicit, and the experimental optimal-effort/equilibrium demonstration replots the analytic maximum of that same payment rule.

  1. self definitional [Section 5.1, Payment Design (Payment_k definition following Theorem 3)]
    "Payment_k(e_k,e'_k) ≜ f( Q/(Φδ_k^2(e_k)+Φδ_k'^2(e'_k)+Υ) ) ∝ 1/Upper Bound(F_c(w_k)-F_c(w_k')) ... the learner can evaluate the value of δ_k and adjust the coefficient Q in the payment function, preventing agents from reducing majority class data samples to attain IID."

    The payment is a decreasing function of the very quantity δ_k that the mechanism claims to incentivize. But δ_k is agent k's private total-variation distance defined in Eq. (1) from p^(k) to p^(c); the server observes only uploaded model updates and the auxiliary set D_c, and no estimation or reporting protocol for δ_k is supplied. Thus the statement that the mechanism rewards agents for reducing their non-IID degree is true by construction: the reward formula is fed the target quantity directly. If the intended payment were instead based on the observable validation gap F_c(w_k)-F_c(w_k'), that would be a different payment rule, and the equilibrium analysis would need to be re-derived. The peer-prediction framing does not repair this, because no elicited response enters the formula.

  2. fitted input called prediction [Section 6.2, Implementing Incentive Mechanism (Eq. (13) and Figure 2)]
    "We can derive and obtain the optimal effort level for logarithmic functions, which can be generally rewritten into a general form δ_k = f(δ_k') = 1/c ± sqrt(1/c^2 - δ_k'^2 - Υ/Φ) ... As shown in Figure 2, the utilities of four randomly selected agents remain stable, which indicates an equilibrium and also confirms Theorem 5 ... the utility achieved by the optimal effort levels is much larger than others in expectation."

    The experiment uses the same utility function defined in Section 5.1 with the same payment formula, the same log scoring function, the same linear cost, and δ_k(e_k)=exp(-e_k). The 'optimal effort level' shown in Eq. (13) is the analytic solution of the first-order condition ∂u_k/∂e_k=0 for that exact utility. Plotting that precomputed optimum and observing that it maximizes the utility is therefore not an empirical finding; it is a restatement of the definition. Similarly, the stability of utilities in Figure 2 follows from evaluating the payment formula with the agents' known, fixed δ values and a randomly selected peer; no independent measurement or out-of-sample validation occurs. The claimed confirmation of equilibrium reduces to the construction of the utility functions.

full rationale

Score 6: the central theorem-level derivations are not circular. Theorem 2 and Example 1 give a genuine bound on the FedAvg divergence term in terms of δ_k, and Theorem 5 is a conditional, standard fixed-point argument showing that well-behaved utilities admit a pure Nash equilibrium; those parts do not reduce to their inputs. The circularity is concentrated at the mechanism level. The payment function in Section 5.1 is defined directly from the private δ_k that the mechanism is supposed to elicit or measure, with no server-visible estimator or reporting protocol, so the incentive effect is definitional rather than derived. The experimental section then 'validates' the mechanism by plugging known δ values into the very payment formula whose analytic optimum was already solved, making the equilibrium/optimality plots true by construction. Self-citations (e.g., to Liu and Wei 2020) appear but are not load-bearing: they motivate peer prediction but the actual scoring rule does not use peer responses. Overall, the paper contains substantial non-circular analysis, but its central implementability claim and the empirical evidence offered for the mechanism reduce by construction, warranting a 6 rather than a lower score.

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

The central mechanism relies on several parameters chosen to make the payment values concrete (L, G, μ, c, Q, the effort-distance function), and on assumptions that are either strong (class-independent Lipschitz constants) or unstated (learner access to private δ_k). No new physical or mathematical entities are introduced beyond the payment mechanism itself.

free parameters (6)
  • L (Lipschitz constant) = 100
    Set in Appendix B.4 using a heuristic based on cross-entropy error rate; directly controls Φ and Υ.
  • G (gradient norm bound) = [0,10]
    Inferred from gradient norms measured on two agents (Appendix B.4); used to set Φ and Υ.
  • μ (strong convexity constant) = 0.01
    Assumed in Appendix B.4 to compute Υ; no justification from the models used.
  • c (marginal cost) = not specified
    Cost coefficient in the cost function; experiments claim a linear cost but the value is not reported.
  • Q (initial payment coefficient) = not specified
    Learner's payment coefficient; no value or selection procedure is specified.
  • δ_k(e_k) effort-distance function = exp(-e_k)
    Chosen in Section 6.2 to follow the Ninety-ninety rule; not derived from data or theory.
assumptions (6)
  • standard math Assumptions 1-4: strong convexity, L-smoothness, bounded local variance, bounded stochastic gradient norm.
    Standard convergence assumptions used in the cited FL literature; invoked in Section 3.1.
  • ad hoc to paper Assumption 5: the payment function f is non-decreasing, differentiable, and strictly concave on [0,1].
    Introduced to ensure a concave utility and a unique best response; the paper argues it is 'mild' and aligned with the Ninety-ninety rule.
  • ad hoc to paper The Lipschitz constant of the loss gradient is identical across all classes: L = L_{x|y=i} for all i.
    Made in Appendix A.2 to simplify the proof of Theorem 2; this is a strong assumption not stated in the main text and not true for typical neural losses.
  • domain assumption The learner can observe or compute each agent's δ_k, the private non-IID degree.
    The payment formula in Section 5.1 depends on δ_k, but the server only receives model updates. No algorithm to recover δ_k from models is provided.
  • domain assumption Agents are rational utility maximizers.
    Underlies the game-theoretic formulation and the equilibrium analysis.
  • domain assumption Effort e_k maps to δ_k through a known non-increasing function.
    The model assumes δ_k(e_k) is known to both agent and learner; in experiments it is set to exp(-e_k).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Incentivizing High-quality Participation From Federated Learning Agents." pith.science (2026). https://pith.science/paper/2QMZ3YNW

@misc{pith2026250616731,
  author       = {Pith},
  title        = {Pith review of: Incentivizing High-quality Participation From Federated Learning Agents},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2QMZ3YNW}},
  note         = {Machine review of arXiv:2506.16731}
}
read the original abstract

Federated learning (FL) provides a promising paradigm for facilitating collaboration between multiple clients that jointly learn a global model without directly sharing their local data. However, existing research suffers from two caveats: 1) From the perspective of agents, voluntary and unselfish participation is often assumed. But self-interested agents may opt out of the system or provide low-quality contributions without proper incentives; 2) From the mechanism designer's perspective, the aggregated models can be unsatisfactory as the existing game-theoretical federated learning approach for data collection ignores the potential heterogeneous effort caused by contributed data. To alleviate above challenges, we propose an incentive-aware framework for agent participation that considers data heterogeneity to accelerate the convergence process. Specifically, we first introduce the notion of Wasserstein distance to explicitly illustrate the heterogeneous effort and reformulate the existing upper bound of convergence. To induce truthful reporting from agents, we analyze and measure the generalization error gap of any two agents by leveraging the peer prediction mechanism to develop score functions. We further present a two-stage Stackelberg game model that formalizes the process and examines the existence of equilibrium. Extensive experiments on real-world datasets demonstrate the effectiveness of our proposed mechanism.

Figures

Figures reproduced from arXiv: 2506.16731 by the authors.

Figure 1
Figure 1. Upper: FL training process under different non-iid degrees; Lower: Performance comparison with peers. 0 50 100 150 round 0 20 40 60 80 100 utility (a) MNIST agent 0 agent 2 agent 3 agent 5 0 50 100 150 round 20 40 60 80 (b) FashionMNIST agent 0 agent 2 agent 3 agent 5 0 100 200 300 round 0 20 40 60 80 100 (c) CIFAR-10 agent 0 agent 2 agent 3 agent 5 0 25 50 75 100 round 30 40 50 60 70 80 (d) CIFAR-100 agent 0 agent … view at source ↗
Figure 2
Figure 2. Upper: Utility variation across different agents; Lower: Utility variation under different non-iid degrees. that is, using the number of classes p within each client’s local dataset. The additional results of this popular non-iid setting (the number of class) are provided in Appendix B.2. 6.2 Implementing Incentive Mechanism Scoring function We first present a scoring structure as the payment function f(·): a logari… view at source ↗
Figure 3
Figure 3. (The number of classes) FL training process under different non-iid degrees. 0 50 100 150 round 20 40 60 80 100 test accuracy (a) MNIST agent 0: p=2 agent 9: p=10 0 50 100 150 round 20 40 60 80 (b) FashionMNIST agent 0: p=2 agent 9: p=10 0 100 200 300 round 20 40 60 (c) CIFAR-10 agent 0: p=2 agent 9: p=10 0 25 50 75 100 round 0 10 20 30 40 (d) CIFAR-100 agent 0: p=4 agent 9: p=20 [PITH_FULL_IMAGE:figures/full_fig_p… view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: (The number of classes) Performance comparison with peers. In default, the non-iid degree is 0.5. B.3 Scoring function settings For linear functions, we have f(ek, ek′ ) = κ Q Φδ 2 k (ek) + Φδ 2 k′ (ek′ ) + Υ [PITH_FULL_IMAGE:figures/full_fig_p022_4.png]
Figure 5
Figure 5. Figure 5: Gradient divergence with peers. The non-iid degree is 0.5. [PITH_FULL_IMAGE:figures/full_fig_p024_5.png]

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

55 extracted references · 35 canonical work pages

  1. [1]

    Programming pearls: algorithm design techniques

    Jon Bentley. Programming pearls: algorithm design techniques. Communications of the ACM , 27(9):865--873, 1984

    work page 1984

  2. [2]

    One for one, or all for all: Equilibria and optimality of collaboration in federated learning

    Avrim Blum, Nika Haghtalab, Richard Lanas Phillips, and Han Shao. One for one, or all for all: Equilibria and optimality of collaboration in federated learning. In International Conference on Machine Learning , pages 1005--1014. PMLR, 2021

    work page 2021

  3. [3]

    \"U ber abbildung von mannigfaltigkeiten

    Luitzen Egbertus Jan Brouwer. \"U ber abbildung von mannigfaltigkeiten. Mathematische annalen , 71(1):97--115, 1911

    work page 1911

  4. [4]

    To federate or not to federate: Incentivizing client participation in federated learning

    Yae Jee Cho, Divyansh Jhunjhunwala, Tian Li, Virginia Smith, and Gauri Joshi. To federate or not to federate: Incentivizing client participation in federated learning. arXiv preprint arXiv:2205.14840 , 2022

    arXiv 2022

  5. [5]

    Maximizing global model appeal in federated learning, 2023

    Yae Jee Cho, Divyansh Jhunjhunwala, Tian Li, Virginia Smith, and Gauri Joshi. Maximizing global model appeal in federated learning, 2023

    work page 2023

  6. [6]

    On the Convergence of Federated Averaging with Cyclic Client Participation

    Yae Jee Cho, Pranay Sharma, Gauri Joshi, Zheng Xu, Satyen Kale, and Tong Zhang. On the convergence of federated averaging with cyclic client participation. arXiv preprint arXiv:2302.03109 , 2023

    work page Pith review arXiv 2023

  7. [7]

    Collaboration equilibrium in federated learning

    Sen Cui, Jian Liang, Weishen Pan, Kun Chen, Changshui Zhang, and Fei Wang. Collaboration equilibrium in federated learning. In Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining , pages 241--251, 2022

    work page 2022

  8. [8]

    Tackling data heterogeneity in federated learning with class prototypes

    Yutong Dai, Zeyuan Chen, Junnan Li, Shelby Heinecke, Lichao Sun, and Ran Xu. Tackling data heterogeneity in federated learning with class prototypes. In Proceedings of the AAAI Conference on Artificial Intelligence , volume 37, pages 7314--7322, 2023

    work page 2023

Show all 55 references
  1. [9]

    Model-sharing games: Analyzing federated learning under voluntary participation

    Kate Donahue and Jon Kleinberg. Model-sharing games: Analyzing federated learning under voluntary participation. In Proceedings of the AAAI Conference on Artificial Intelligence , 2021

    2021

  2. [10]

    Optimality and stability in federated learning: A game-theoretic approach

    Kate Donahue and Jon Kleinberg. Optimality and stability in federated learning: A game-theoretic approach. Advances in Neural Information Processing Systems , 34:1287--1298, 2021

    2021

  3. [11]

    Online scheduling unbiased distributed learning over wireless edge networks

    Ziyi Han, Ruiting Zhou, Jinlong Pang, Yue Cao, and Haisheng Tan. Online scheduling unbiased distributed learning over wireless edge networks. In 2021 IEEE 27th International Conference on Parallel and Distributed Systems (ICPADS) , pages 599--606. IEEE, 2021

    2021

  4. [12]

    Tokenized incentive for federated learning

    Jingoo Han, Ahmad Faraz Khan, Syed Zawad, Ali Anwar, Nathalie Baracaldo Angel, Yi Zhou, Feng Yan, and Ali R Butt. Tokenized incentive for federated learning. In Proceedings of the Federated Learning Workshop at the Association for the Advancement of Artificial Intelligence (AA...

    2022

  5. [13]

    Incentive mechanism design for federated learning: Hedonic game approach

    Cengis Hasan. Incentive mechanism design for federated learning: Hedonic game approach. arXiv preprint arXiv:2101.09673 , 2021

    2021 arXiv

  6. [14]

    Towards practical overlay networks for decentralized federated learning

    Yifan Hua, Jinlong Pang, Xiaoxue Zhang, Yi Liu, Xiaofeng Shi, Bao Wang, Yang Liu, and Chen Qian. Towards practical overlay networks for decentralized federated learning. In 2024 IEEE 32nd International Conference on Network Protocols (ICNP) , pages 1--11. IEEE, 2024

    2024

  7. [15]

    Batch normalization: Accelerating deep network training by reducing internal covariate shift

    Sergey Ioffe and Christian Szegedy. Batch normalization: Accelerating deep network training by reducing internal covariate shift. In International conference on machine learning , pages 448--456. pmlr, 2015

    2015

  8. [16]

    Incentive design for efficient federated learning in mobile networks: A contract theory approach

    Jiawen Kang, Zehui Xiong, Dusit Niyato, Han Yu, Ying-Chang Liang, and Dong In Kim. Incentive design for efficient federated learning in mobile networks: A contract theory approach. In 2019 IEEE VTS Asia Pacific Wireless Communications Symposium (APWCS) , pages 1--5. IEEE, 2019

    2019

  9. [17]

    Scaffold: Stochastic controlled averaging for federated learning

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

    2020

  10. [18]

    Mechanisms that incentivize data sharing in federated learning

    Sai Praneeth Karimireddy, Wenshuo Guo, and Michael I Jordan. Mechanisms that incentivize data sharing in federated learning. arXiv preprint arXiv:2207.04557 , 2022

    2022 arXiv

  11. [19]

    Federated learning for edge networks: Resource optimization and incentive mechanism

    Latif U Khan, Shashi Raj Pandey, Nguyen H Tran, Walid Saad, Zhu Han, Minh NH Nguyen, and Choong Seon Hong. Federated learning for edge networks: Resource optimization and incentive mechanism. IEEE Communications Magazine , 58(10):88--93, 2020

    2020

  12. [20]

    Adam: A method for stochastic optimization

    Diederik P Kingma and Jimmy Ba. Adam: A method for stochastic optimization. International Conference on Learning Representations , 2015

    2015

  13. [21]

    A unified theory of decentralized sgd with changing topology and local updates

    Anastasia Koloskova, Nicolas Loizou, Sadra Boreiri, Martin Jaggi, and Sebastian Stich. A unified theory of decentralized sgd with changing topology and local updates. In International Conference on Machine Learning , pages 5381--5393. PMLR, 2020

    2020

  14. [22]

    Incentivizing federated learning, 2022

    Shuyu Kong, You Li, and Hai Zhou. Incentivizing federated learning, 2022

    2022

  15. [23]

    Learning multiple layers of features from tiny images

    Alex Krizhevsky, Geoffrey Hinton, et al. Learning multiple layers of features from tiny images. 2009

    2009

  16. [24]

    Gradient-based learning applied to document recognition

    Yann LeCun, L \'e on Bottou, Yoshua Bengio, and Patrick Haffner. Gradient-based learning applied to document recognition. Proceedings of the IEEE , 86(11):2278--2324, 1998

    1998

  17. [25]

    On the convergence of fedavg on non-iid data

    Xiang Li, Kaixuan Huang, Wenhao Yang, Shusen Wang, and Zhihua Zhang. On the convergence of fedavg on non-iid data. In International Conference on Learning Representations , 2019

    2019

  18. [26]

    Federated optimization in heterogeneous networks

    Tian Li, Anit Kumar Sahu, Manzil Zaheer, Maziar Sanjabi, Ameet Talwalkar, and Virginia Smith. Federated optimization in heterogeneous networks. Proceedings of Machine learning and systems , 2:429--450, 2020

    2020

  19. [27]

    Incentives for federated learning: a hypothesis elicitation approach

    Yang Liu and Jiaheng Wei. Incentives for federated learning: a hypothesis elicitation approach. arXiv preprint arXiv:2007.10596 , 2020

    2007 arXiv

  20. [28]

    Automatic dataset construction (adc): Sample collection, data curation, and beyond

    Minghao Liu, Zonglin Di, Jiaheng Wei, Zhongruo Wang, Hengxiang Zhang, Ruixuan Xiao, Haoyu Wang, Jinlong Pang, Hao Chen, Ankit Shah, et al. Automatic dataset construction (adc): Sample collection, data curation, and beyond. arXiv preprint arXiv:2408.11338 , 2024

    2024 arXiv

  21. [29]

    Communication-efficient learning of deep networks from decentralized data

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

    2017

  22. [30]

    Eliciting informative feedback: The peer-prediction method

    Nolan Miller, Paul Resnick, and Richard Zeckhauser. Eliciting informative feedback: The peer-prediction method. Management Science , 51(9):1359--1373, 2005

    2005

  23. [31]

    Online scheduling algorithms for unbiased distributed learning over wireless edge networks

    Jinlong Pang, Ziyi Han, Ruiting Zhou, Haisheng Tan, and Yue Cao. Online scheduling algorithms for unbiased distributed learning over wireless edge networks. Journal of Systems Architecture , 131:102673, 2022

    2022

  24. [32]

    An incentive auction for heterogeneous client selection in federated learning

    Jinlong Pang, Jieling Yu, Ruiting Zhou, and John CS Lui. An incentive auction for heterogeneous client selection in federated learning. IEEE Transactions on Mobile Computing , 2022

    2022

  25. [33]

    Eris: An online auction for scheduling unbiased distributed learning over edge networks

    Jinlong Pang, Ziyi Han, Ruiting Zhou, Renli Zhang, John CS Lui, and Hao Chen. Eris: An online auction for scheduling unbiased distributed learning over edge networks. IEEE Transactions on Mobile Computing , 23(6):7196--7209, 2023

    2023

  26. [34]

    Fairness without harm: An influence-guided active sampling approach

    Jinlong Pang, Jialu Wang, Zhaowei Zhu, Yuanshun Yao, Chen Qian, and Yang Liu. Fairness without harm: An influence-guided active sampling approach. Advances in Neural Information Processing Systems , 37:61513--61548, 2024

    2024

  27. [35]

    Improving data efficiency via curating llm-driven rating systems

    Jinlong Pang, Jiaheng Wei, Ankit Parag Shah, Zhaowei Zhu, Yaxuan Wang, Chen Qian, Yang Liu, Yujia Bao, and Wei Wei. Improving data efficiency via curating llm-driven rating systems. arXiv preprint arXiv:2410.10877 , 2024

    2024 arXiv

  28. [36]

    Token cleaning: Fine-grained data selection for llm supervised fine-tuning

    Jinlong Pang, Na Di, Zhaowei Zhu, Jiaheng Wei, Hao Cheng, Chen Qian, and Yang Liu. Token cleaning: Fine-grained data selection for llm supervised fine-tuning. arXiv preprint arXiv:2502.01968 , 2025

    2025

  29. [37]

    Making gradient descent optimal for strongly convex stochastic optimization

    Alexander Rakhlin, Ohad Shamir, and Karthik Sridharan. Making gradient descent optimal for strongly convex stochastic optimization. arXiv preprint arXiv:1109.5647 , 2011

    2011 arXiv

  30. [38]

    Informed truthfulness in multi-task peer prediction

    Victor Shnayder, Arpit Agarwal, Rafael Frongillo, and David C Parkes. Informed truthfulness in multi-task peer prediction. In Proceedings of the 2016 ACM Conference on Economics and Computation , pages 179--196, 2016

    2016

  31. [39]

    Local sgd converges fast and communicates little

    Sebastian U Stich. Local sgd converges fast and communicates little. arXiv preprint arXiv:1805.09767 , 2018

    2018 arXiv

  32. [40]

    Scipy 1.0: fundamental algorithms for scientific computing in python

    Pauli Virtanen, Ralf Gommers, Travis E Oliphant, Matt Haberland, Tyler Reddy, David Cournapeau, Evgeni Burovski, Pearu Peterson, Warren Weckesser, Jonathan Bright, et al. Scipy 1.0: fundamental algorithms for scientific computing in python. Nature methods , 17(3):261--272, 2020

    2020

  33. [41]

    Adaptive federated learning in resource constrained edge computing systems

    Shiqiang Wang, Tiffany Tuor, Theodoros Salonidis, Kin K Leung, Christian Makaya, Ting He, and Kevin Chan. Adaptive federated learning in resource constrained edge computing systems. IEEE Journal on Selected Areas in Communications , 37(6):1205--1221, 2019

    2019

  34. [42]

    A principled approach to data valuation for federated learning

    Tianhao Wang, Johannes Rausch, Ce Zhang, Ruoxi Jia, and Dawn Song. A principled approach to data valuation for federated learning. Federated Learning: Privacy and Incentive , pages 153--167, 2020

    2020

  35. [43]

    Sample elicitation

    Jiaheng Wei, Zuyue Fu, Yang Liu, Xingyu Li, Zhuoran Yang, and Zhaoran Wang. Sample elicitation. In International Conference on Artificial Intelligence and Statistics , pages 2692--2700. PMLR, 2021

    2021

  36. [44]

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

    Han Xiao, Kashif Rasul, and Roland Vollgraf. Fashion-mnist: a novel image dataset for benchmarking machine learning algorithms. arXiv preprint arXiv:1708.07747 , 2017

    2017 arXiv

  37. [45]

    Gradient driven rewards to guarantee fairness in collaborative machine learning

    Xinyi Xu, Lingjuan Lyu, Xingjun Ma, Chenglin Miao, Chuan Sheng Foo, and Bryan Kian Hsiang Low. Gradient driven rewards to guarantee fairness in collaborative machine learning. Advances in Neural Information Processing Systems , 34:16104--16117, 2021

    2021

  38. [46]

    Achieving linear speedup with partial worker participation in non-iid federated learning

    Haibo Yang, Minghong Fang, and Jia Liu. Achieving linear speedup with partial worker participation in non-iid federated learning. arXiv preprint arXiv:2101.11203 , 2021

    2021 arXiv

  39. [47]

    Anarchic federated learning

    Haibo Yang, Xin Zhang, Prashant Khanduri, and Jia Liu. Anarchic federated learning. In International Conference on Machine Learning , pages 25331--25363. PMLR, 2022

    2022

  40. [48]

    Parallel restarted sgd with faster convergence and less communication: Demystifying why model averaging works for deep learning

    Hao Yu, Sen Yang, and Shenghuo Zhu. Parallel restarted sgd with faster convergence and less communication: Demystifying why model averaging works for deep learning. In Proceedings of the AAAI Conference on Artificial Intelligence , 2019

    2019

  41. [49]

    A fairness-aware incentive scheme for federated learning

    Han Yu, Zelei Liu, Yang Liu, Tianjian Chen, Mingshu Cong, Xi Weng, Dusit Niyato, and Qiang Yang. A fairness-aware incentive scheme for federated learning. In Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society , pages 393--399, 2020

    2020

  42. [50]

    Faithful edge federated learning: Scalability and privacy

    Meng Zhang, Ermin Wei, and Randall Berry. Faithful edge federated learning: Scalability and privacy. IEEE Journal on Selected Areas in Communications , 39(12):3790--3804, 2021

    2021

  43. [51]

    Evaluating llm-corrupted crowdsourcing data without ground truth

    Yichi Zhang, Jinlong Pang, Zhaowei Zhu, and Yang Liu. Evaluating llm-corrupted crowdsourcing data without ground truth. arXiv preprint arXiv:2506.06991 , 2025

    2025

  44. [52]

    Federated learning with non-iid data

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

    2018 arXiv

  45. [53]

    A truthful procurement auction for incentivizing heterogeneous clients in federated learning

    Ruiting Zhou, Jinlong Pang, Zhibo Wang, John CS Lui, and Zongpeng Li. A truthful procurement auction for incentivizing heterogeneous clients in federated learning. In 2021 IEEE 41st International Conference on Distributed Computing Systems (ICDCS) , pages 183--193. IEEE, 2021

    2021

  46. [54]

    Online scheduling algorithm for heterogeneous distributed machine learning jobs

    Ruiting Zhou, Jinlong Pang, Qin Zhang, Chuan Wu, Lei Jiao, Yi Zhong, and Zongpeng Li. Online scheduling algorithm for heterogeneous distributed machine learning jobs. IEEE Transactions on Cloud Computing , 11(2):1514--1529, 2022

    2022

  47. [55]

    write newline

    " write newline "" before.all 'output.state := FUNCTION fin.entry add.period write newline FUNCTION new.block output.state before.all = 'skip after.block 'output.state := if FUNCTION new.sentence output.state after.block = 'skip output.state before.all = 'skip after.sentence '...

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.