AI Paper Atlas

Research map / Federated learning and privacy / Communication-efficient FL

Re-architecting Personalized Federated Learning for Demanding Edge Environments

Quyang Pan, Sheng Sun, Tingting Wi, Zhiyuan Wu, Yuwei Wang, Min Liu, Bo Gao, Jingyuan Wang

AAAI 2026 · Federated learning and privacy › Communication-efficient FL · approach: Efficient local training in FL · 0 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
replace logits with distilled data as cached knowledge (DistilCacheFL vs FedCache)knowledge cache-driven FL · CIFAR-10 (alpha=0.5, homogeneous models)average user accuracy (%): 42.1 -> 51.1Improved
replace logits with distilled data as cached knowledge (DistilCacheFL vs FedCache)knowledge cache-driven FL · CIFAR-100 (alpha=2, homogeneous models)average user accuracy (%): 14.7 -> 23.3Improved
increase cache sampling tau from 0 to 0.3 (ablation)DistilCacheFL · CIFAR-10 (alpha=0.5, homogeneous models)average user accuracy (%): 51.3 -> 51.7Improved
increase cache sampling tau toward 1.0 (sample whole cache) (ablation)DistilCacheFL · CIFAR-10 (alpha=0.5, homogeneous models)average user accuracy (%): 51.7 -> 48.6Hurt
use heterogeneous ResNet-S/M/L models instead of homogeneous ResNet-S (ablation)DistilCacheFL · CIFAR-10 (alpha=0.5)average user accuracy (%): 46.6 -> 51.1Improved
use heterogeneous ResNet-S/M/L models instead of homogeneous ResNet-S (ablation)DistilCacheFL · CIFAR-100 (alpha=0.5)average user accuracy (%): 31.5 -> 35.8Improved
use heterogeneous ResNet-S/M/L models instead of homogeneous ResNet-S (ablation)FedCache · CIFAR-10 (alpha=0.5)average user accuracy (%): 40.5 -> 41.3Improved
model heterogeneity vs homogeneity in the baselines FedKD and FedCacheFedCache, FedKD · CIFAR-10 (alpha=2)average user accuracy (%): FedCache 23.9 -> 22.2No clear effect

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