Pith. sign in

REVIEW 3 cited by

On the Unreasonable Effectiveness of Federated Averaging with Heterogeneous Data

Not yet reviewed by Pith; the record is open.

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 2206.04723 v1 pith:YNAKXV2R submitted 2022-06-09 cs.LG

On the Unreasonable Effectiveness of Federated Averaging with Heterogeneous Data

classification cs.LG
keywords fedavgdatafederateddissimilaritygradientheterogeneitytheoreticalalgorithm
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Existing theory predicts that data heterogeneity will degrade the performance of the Federated Averaging (FedAvg) algorithm in federated learning. However, in practice, the simple FedAvg algorithm converges very well. This paper explains the seemingly unreasonable effectiveness of FedAvg that contradicts the previous theoretical predictions. We find that the key assumption of bounded gradient dissimilarity in previous theoretical analyses is too pessimistic to characterize data heterogeneity in practical applications. For a simple quadratic problem, we demonstrate there exist regimes where large gradient dissimilarity does not have any negative impact on the convergence of FedAvg. Motivated by this observation, we propose a new quantity, average drift at optimum, to measure the effects of data heterogeneity, and explicitly use it to present a new theoretical analysis of FedAvg. We show that the average drift at optimum is nearly zero across many real-world federated training tasks, whereas the gradient dissimilarity can be large. And our new analysis suggests FedAvg can have identical convergence rates in homogeneous and heterogeneous data settings, and hence, leads to better understanding of its empirical success.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 3 Pith papers

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

  1. What's in a Smoothness Constant? Tighter Rates for Local SGD with Bounded Second-order Heterogeneity

    cs.LG 2026-07 conditional novelty 7.0

    Local SGD provably improves over Mini-batch SGD under bounded second-order heterogeneity in the general convex setting, with nearly tight upper and lower bounds.

  2. On the Convergence Rates of Federated Q-Learning across Heterogeneous Environments

    cs.LG 2024-09 unverdicted novelty 7.0

    Federated Q-learning in heterogeneous environments achieves linear speedup in K agents for sampling error but is limited to Θ(E/T) convergence when averaging every E steps, with a two-phase error decay-then-rise behav...

  3. How to DP-fy Your Data: A Practical Guide to Generating Synthetic Data With Differential Privacy

    cs.CR 2025-12 conditional novelty 2.0

    A practical, extremely thorough survey of differentially private synthetic data generation: methods, privacy units, evaluation metrics, and end-to-end system components across four data modalities.