Pith. sign in

REVIEW 53 cited by

LEAF: A Benchmark for Federated Settings

Not yet reviewed by Pith; the record is open.

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

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 1812.01097 v3 pith:4YPAYJIS submitted 2018-12-03 cs.LG stat.ML

Sebastian Caldas , Sai Meher Karthik Duddu , Peter Wu , Tian Li , Jakub Konečný , H. Brendan McMahan , Virginia Smith , Ameet Talwalkar This is my paper · ORCID
classification cs.LGstat.ML
keywords federatedlearningdataleafareaschallengesframeworksettings
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Modern federated networks, such as those comprised of wearable devices, mobile phones, or autonomous vehicles, generate massive amounts of data each day. This wealth of data can help to learn models that can improve the user experience on each device. However, the scale and heterogeneity of federated data presents new challenges in research areas such as federated learning, meta-learning, and multi-task learning. As the machine learning community begins to tackle these challenges, we are at a critical time to ensure that developments made in these areas are grounded with realistic benchmarks. To this end, we propose LEAF, a modular benchmarking framework for learning in federated settings. LEAF includes a suite of open-source federated datasets, a rigorous evaluation framework, and a set of reference implementations, all geared towards capturing the obstacles and intricacies of practical federated environments.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 53 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. OpenAlex reports about 285 citations worldwide. Full citation record

  1. SpecGradFilter: A Spectral Gradient Filtering Framework for Taming Federated Heterogeneity

    cs.LG 2026-07 conditional novelty 7.0 of 10

    Inter-client gradient divergence in federated learning concentrates in low-frequency components; suppressing them via spectral or spatial high-pass filtering reduces client drift and raises accuracy under non-IID data.

  2. FedFFT: Taming Client Drift in Federated SAM via Spectral Perturbation Filtering

    cs.LG 2026-07 conditional novelty 6.5 of 10

    Low-frequency components of client-side SAM perturbations carry most inter-client disagreement; high-pass filtering them yields more consistent federated updates and higher accuracy under non-IID data.

  3. pFedUL: Layer-Aware Federated Unlearning for Personalized Federated Learning

    cs.LG 2026-06 conditional novelty 6.5 of 10

    Layer-aware selective unlearning for personalized FL matches near-retrain forgetting while retaining ~97% personalized accuracy for remaining clients across four pFL architectures.

  4. Capacity Confounds and Coverage Guarantees in Adaptive Sub-model Federated Learning

    cs.LG 2026-08 conditional novelty 6.0 of 10

    Update-based estimates of client data heterogeneity in sub-model federated learning are dominated by device capacity, and adaptive allocation adds nothing over a matched-budget random control once parameter coverage i...

  5. AutoEncoder-Compressed Parallel Split Learning for Pre-trained Model Fine-Tuning

    cs.DC 2026-07 conditional novelty 6.0 of 10

    An autoencoder-based split-learning compressor with a two-stage alignment protocol achieves about 10x communication reduction during pre-trained vision-model fine-tuning with near-zero accuracy loss, outperforming heu...

  6. NFSA: Non-Forward Secure Aggregation with One Server via Two Layer Secret Sharing

    cs.CR 2026-07 conditional novelty 6.0 of 10

    NFSA combines PRF-based two-layer secret sharing with CRT packing of almost key-homomorphic PRF masks to achieve single-server, one-shot secure aggregation without server-forwarded key shares.

  7. PRoVeFL: Private Robust and Verifiable Aggregation in Federated Learning

    cs.CR 2026-07 conditional novelty 6.0 of 10

    Multi-server multi-key FHE with a shared random mask lets PRoVeFL run complex Byzantine-robust FL aggregation privately and verifiably, with large reported speedups over Prio and ELSA.

  8. Benchmarking Robust Aggregation in Decentralized Gradient Marketplaces

    cs.LG 2025-09 conditional novelty 6.0 of 10

    Adaptive Sybil backdoor attacks can defeat MartFL, FLTrust, and SkyMask in buyer-baseline gradient marketplaces with little visible effect on accuracy or cost.

  9. PracMHBench: Re-evaluating Model-Heterogeneous Federated Learning Based on Practical Edge Device Constraints

    cs.LG 2025-09 conditional novelty 6.0 of 10

    PracMHBench evaluates eight model-heterogeneous federated learning algorithms under practical edge device constraints and finds that depth-level heterogeneity wins under compute/communication limits while memory limit...

  10. Differentially Private Federated Clustering with Random Rebalancing

    cs.LG 2025-08 reject novelty 6.0 of 10

    RR-Cluster enforces a minimum cluster size by random rebalancing, lowering DP noise and improving federated clustering utility, but its privacy proof understates the true noise.

  11. Flotilla: A scalable, modular and resilient federated learning framework for heterogeneous resources

    cs.DC 2025-07 conditional novelty 6.0 of 10

    Flotilla is a modular, resilient federated learning framework that runs on heterogeneous edge devices, supports sync and async strategies, and scales to 1000+ clients with low overhead.

  12. FedAPM: Federated Learning via ADMM with Partial Model Personalization

    cs.LG 2025-06 conditional novelty 6.0 of 10

    FedAPM applies ADMM with first- and second-order proximal corrections to partial model personalization in federated learning, proving global convergence and reporting better accuracy, F1, and AUC than FedAlt, FedSim, ...

  13. HALoS: Hierarchical Asynchronous Local SGD over Slow Networks for Geo-Distributed Large Language Model Training

    cs.LG 2025-06 conditional novelty 6.0 of 10

    A hierarchical asynchronous local SGD method with regional parameter servers and global model merging is claimed to train small LLMs up to 7.5x faster than DiLoCo in simulated geo-distributed settings.

  14. HtFLlib: A Comprehensive Heterogeneous Federated Learning Library and Benchmark

    cs.LG 2025-06 conditional novelty 6.0 of 10

    HtFLlib is a unified benchmark and library with 12 datasets, 40 heterogeneous model architectures, and systematic accuracy, convergence, and cost evaluations of 10 HtFL methods.

  15. DRAUN: An Algorithm-Agnostic Data Reconstruction Attack on Federated Unlearning Systems

    cs.LG 2025-06 conditional novelty 6.0 of 10

    DRAUN reconstructs unlearned client images from federated unlearning updates by simulating possible unlearning losses and matching gradients, exposing privacy leakage in optimization-based federated unlearning.

  16. Generalized and Personalized Federated Learning with Black-Box Foundation Models via Orthogonal Transformations

    cs.LG 2025-05 conditional novelty 6.0 of 10

    A federated learning method combines client-specific orthogonal transformations on frozen black-box foundation model embeddings with a shared classifier, outperforming baselines on several domain-shift benchmarks.

  17. Exploit Gradient Skewness to Circumvent Byzantine Defenses for Federated Learning

    cs.LG 2025-02 conditional novelty 6.0 of 10

    A skew-aware Byzantine attack, STRIKE, exploits the tendency of honest non-IID gradients to form dense clusters away from their mean, hiding malicious gradients inside the cluster.

  18. Aequa: Fair Model Rewards in Collaborative Learning via Slimmable Networks

    cs.LG 2025-02 conditional novelty 6.0 of 10

    Aequa allocates model widths (and thus accuracies) to federated learning participants in proportion to their estimated contributions, using slimmable networks and a simulated annealing optimizer.

  19. Decoding FL Defenses: Systemization, Pitfalls, and Remedies

    cs.CR 2025-02 conditional novelty 6.0 of 10

    Many FL defenses are evaluated on overly easy datasets and attacks, and this paper demonstrates with case studies that those easy settings can make weak defenses look strong.

  20. THOR: A Generic Energy Estimation Approach for On-Device Training

    cs.LG 2025-01 conditional novelty 6.0 of 10

    A layer-wise Gaussian Process model estimates DNN training energy from measured probe networks, reducing MAPE from about 40% to about 10% versus FLOPs-based estimation.

  21. Personalized Language Model Learning on Text Data Without User Identifiers

    cs.LG 2025-01 reject novelty 6.0 of 10

    IDfree-PL samples user-specific embedding distributions on-device to personalize cloud language models without explicit user IDs, with theoretical and empirical support for accuracy gains and embedding-attribution resistance.

  22. FedCFA: Alleviating Simpson's Paradox in Model Aggregation with Counterfactual Federated Learning

    cs.LG 2024-12 conditional novelty 6.0 of 10

    FedCFA replaces local latent factors with global average features to generate counterfactual samples, improving federated global model accuracy under heterogeneous data.

  23. Incentivizing Truthful Collaboration in Heterogeneous Federated Learning

    cs.LG 2024-12 conditional novelty 6.0 of 10

    A norm-comparison payment rule makes truthful gradient reporting an approximately optimal strategy for clients in heterogeneous federated learning.

  24. Non-IID data in Federated Learning: A Survey with Taxonomy, Metrics, Methods, Frameworks and Future Directions

    cs.LG 2024-11 conditional novelty 6.0 of 10

    A comprehensive survey that organizes non-IID data in federated learning into taxonomies of skew types, partition protocols, and metrics, with a meta-analysis of 235 selected papers.

  25. HERO: A Heterogeneity-Aware Benchmark Library for Federated Continual Learning

    cs.LG 2026-06 conditional novelty 5.5 of 10

    HERO shows that FCL method rankings shift when client data skew and task-order mismatch are controlled separately, and that average accuracy can hide weak bottom-client performance.

  26. Targeted Label-Flipping and Oversampling Attacks on Federated Conditional GANs

    cs.LG 2026-08 conditional novelty 5.0 of 10

    Label-flipping and oversampling attacks on federated conditional GANs shift the generated target-class distribution toward the source class linearly in poisoning strength while only quadratically changing the true tar...

  27. Reputation-driven Cooperation in Lattice-based Decentralized Federated Learning through Evolutionary Game Theory

    cs.AI 2026-08 reject novelty 5.0 of 10

    In a lattice-based simulation of decentralized federated learning, a reputation mechanism that rewards cooperators and penalizes defectors raises average accuracy from 70% to 82% and drives cooperation to near 100%.

  28. DFCA: Decentralized Federated Clustering Algorithm

    cs.LG 2025-10 conditional novelty 5.0 of 10

    DFCA decentralizes IFCA-style clustered federated learning: clients keep one model per cluster, train their assigned model locally, and exchange only that model with neighbors via a running average, matching centraliz...

  29. FLAegis: A Two-Layer Defense Framework for Federated Learning Against Poisoning Attacks

    cs.LG 2025-08 conditional novelty 5.0 of 10

    FLAegis defends federated learning by SAX-transforming client updates, spectral-clustering them to filter malicious clients, and applying FFT-based robust aggregation, outperforming several baselines on FEMNIST.

  30. FedBKD: Distilled Federated Learning to Embrace Gerneralization and Personalization on Non-IID Data

    cs.LG 2025-06 conditional novelty 5.0 of 10

    A data-free GAN plus bidirectional knowledge distillation between global and local models improves both personalization and generalization in non-IID federated classification.

  31. Tackling Heterogeneity in Federated Learning via Variance-Reduced Boltzmann Sampling within Homogeneous Social Coalitions

    cs.LG 2025-06 reject novelty 5.0 of 10

    A variance-reduction-based client selection with coalition clustering yields modest accuracy gains over baselines in heterogeneous federated learning, but its convergence guarantee rests on an assumption that the poli...

  32. Federated Learning with Unlabeled Clients: Personalization Can Happen in Low Dimensions

    cs.LG 2025-05 conditional novelty 5.0 of 10

    FLowDUP generates personalized federated models for unlabeled clients via a hypernetwork operating in a low-dimensional random subspace, with a transductive multi-task PAC-Bayes bound motivating the objective.

  33. PLayer-FL: A Principled Approach to Personalized Layer-wise Cross-Silo Federated Learning

    cs.LG 2025-02 conditional novelty 5.0 of 10

    PLayer-FL picks the layer split in partial federated learning from a low-cost sensitivity metric computed at epoch 1, and reports competitive F1, fairness, and participation incentives across seven non-IID datasets.

  34. Interaction-Aware Gaussian Weighting for Clustered Federated Learning

    cs.LG 2025-02 conditional novelty 5.0 of 10

    A loss-trajectory-based clustering method for federated learning, plus a Wasserstein-adjusted cluster metric, reports improved personalization on heterogeneous data.

  35. Distributed Quasi-Newton Method for Fair and Fast Federated Learning

    cs.LG 2025-01 reject novelty 5.0 of 10

    DQN-Fed updates a global model in a direction that makes every client's loss decrease at a rate tied to its local quasi-Newton step, with claimed linear-quadratic convergence.

  36. Benchmarking Federated Learning for Semantic Datasets: Federated Scene Graph Generation

    cs.CV 2024-12 conditional novelty 5.0 of 10

    A clustering-based process for creating federated learning benchmarks with controllable semantic heterogeneity, demonstrated on panoptic scene graph generation and CelebA.

  37. Distributed, communication-efficient, and differentially private estimation of KL divergence

    cs.LG 2024-11 reject novelty 5.0 of 10

    PRIEST-KLD is a family of differentially private, communication-efficient estimators of KL divergence for federated data, with three trust models; however, the unbiasedness and privacy proofs have load-bearing gaps.

  38. Partial Knowledge Distillation for Alleviating the Inherent Inter-Class Discrepancy in Federated Learning

    cs.LG 2024-11 conditional novelty 5.0 of 10

    Weak classes that are intrinsically confusable persist under class-balanced federated learning; a partial knowledge distillation method that distills from class-specific experts improves their accuracy.

  39. Towards Effective Device-Aware Federated Learning

    cs.LG 2019-08 conditional novelty 5.0 of 10

    Federated learning aggregation can be improved by weighting clients with label diversity and model divergence, but tuning the priority order on test accuracy overstates the benefit.

  40. Federated Learning for Object Detection: Enabling Collaborative Drone Learning Without Centralizing Data

    cs.LG 2026-07 conditional novelty 4.0 of 10

    FedAvg on non-IID KIIT-MiTA drone imagery recovers most centralized YOLO nano mAP while keeping images local, with YOLO26 nano gaining ~53% and ~68% relative mAP over single-drone baselines.

  41. PrivacyBench: Privacy Isn't Free in Hybrid Privacy-Preserving Vision Systems

    cs.CR 2026-02 conditional novelty 4.0 of 10

    Combining federated learning with differential privacy causes catastrophic accuracy loss and large resource overhead in vision models, whereas federated learning with secure multi-party computation retains near-baseli...

  42. FedEve: On Bridging the Client Drift and Period Drift for Cross-device Federated Learning

    cs.LG 2025-08 reject novelty 4.0 of 10

    FedEve uses a Kalman filter to combine server momentum (prediction) with client updates (observation) to offset period drift and client drift in cross-device federated learning.

  43. Collaborative Inference and Learning between Edge SLMs and Cloud LLMs: A Survey of Algorithms, Execution, and Open Challenges

    cs.DC 2025-07 conditional novelty 4.0 of 10

    A survey that builds a taxonomy of edge-cloud LLM-SLM collaboration for inference and training, claiming to be the first to unify both phases.

  44. Federated Split Learning with Improved Communication and Storage Efficiency

    cs.LG 2025-07 conditional novelty 4.0 of 10

    CSE-FSL combines an auxiliary network for local updates with periodic smashed-data uploads and a single server-side model, claiming convergence under non-convex loss and lower communication and storage costs.

  45. Evaluating the Impact of Privacy-Preserving Federated Learning on CAN Intrusion Detection

    cs.CR 2025-06 conditional novelty 4.0 of 10

    A federated FedAvg implementation of the CANdito LSTM autoencoder IDS achieves usable detection rates on the ReCAN dataset with slightly lower performance but higher communication cost than centralized training.

  46. Efficient Data Labeling and Optimal Device Scheduling in HWNs Using Clustered Federated Semi-Supervised Learning

    cs.DC 2024-12 conditional novelty 4.0 of 10

    A clustered federated learning framework uses specialized or ensemble models to pseudo-label unlabeled device data, with timing and scheduling heuristics that the authors say improve accuracy and cut energy use.

  47. Hybrid-Regularized Magnitude Pruning for Robust Federated Learning under Covariate Shift

    cs.LG 2024-12 reject novelty 4.0 of 10

    FEDMPR, combining magnitude pruning, dropout, and noise injection in local training, reports accuracy gains over standard federated baselines on several image benchmarks, though not consistently in all settings.

  48. PIcsC: Partitioning-Induced Covariate Shift Correction

    cs.LG 2026-07 reject novelty 3.0 of 10

    A Fisher-information regularizer is proposed to correct partition-induced covariate shift in cross-validation and federated learning, with reported gains of 3-5 points over FedAvg-class baselines.

  49. Strategies for Improving Communication Efficiency in Distributed and Federated Learning: Compression, Local Training, and Personalization

    cs.LG 2025-09 conditional novelty 3.0 of 10

    A PhD dissertation showing unified compression theory, personalized accelerated local training, and pruning methods that reduce communication costs in federated learning and maintain accuracy in LLM pruning.

  50. Accelerated Training of Federated Learning via Second-Order Methods

    cs.LG 2025-05 conditional novelty 3.0 of 10

    A survey that categorizes second-order federated learning methods and argues they reduce communication rounds, based on results borrowed from the cited papers rather than new experiments.

  51. fluke: Federated Learning Utility frameworK for Experimentation and research

    cs.LG 2024-12 conditional novelty 3.0 of 10

    fluke is an open-source Python package that simulates federated learning locally, letting researchers prototype new FL algorithms by defining just client and server behaviors.

  52. Federated Learning with Additional Mechanisms on Clients to Reduce Communication Costs

    cs.LG 2019-08 conditional novelty 3.0 of 10

    Adding MMD regularization or feature fusion modules to on-device training can reduce federated learning communication rounds by 20-60 percent on MNIST and CIFAR-10, according to the authors' experiments.

  53. Federated Learning: Challenges, Methods, and Future Directions

    cs.LG 2019-08 unverdicted

    This survey maps federated learning's core challenges, reviews existing methods, and lists open problems.

Pith tools