AI Paper Atlas

Research map / Federated learning and privacy / Personalized and heterogeneous FL

One-Pass Distribution Sketch for Measuring Data Heterogeneity in Federated Learning

Zichang Liu, Zhaozhuo Xu, Benjamin Coleman, Anshumali Shrivastava

NeurIPS 2023 · Federated learning and privacy › Personalized and heterogeneous FL · approach: Federated learning with pretrained models · 5 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
retrieve the nearest-neighbor personalized model via sketch distance instead of the global modelcold-start personalized FL · MNIST cold clientstest accuracy: 0.8791 -> 0.9702Improved
retrieve the nearest-neighbor personalized model instead of the global modelcold-start personalized FL · FEMNIST cold clientstest accuracy: 0.6667 -> 0.6937Improved
retrieve the nearest-neighbor personalized model instead of the global modelcold-start personalized FL · Shakespeare cold clientstest accuracy: 0.3237 -> 0.3455Improved
local fine-tuning of global model for warm clientsFedAvg global model · MNIST warm clientstest accuracy: 0.8764 -> 0.9784Improved
local fine-tuning of global model for warm clientsFedAvg global model · FEMNIST warm clientstest accuracy: 0.6593 -> 0.7336Improved
local fine-tuning of global model for warm clientsFedAvg global model · Shakespeare warm clientstest accuracy: 0.3581 -> 0.3645No clear effect
sketch-distance client selection instead of uniform random samplingFedAvg · MNIST, MNIST Uniform+Dirichlet, FEMNIST, Shakespearetest accuracy (convergence)Improved
use sketch-based neighbor model instead of random personalized model (ablation)cold-start personalized FL · MNIST cold clientstest accuracyImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy