AI Paper Atlas

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

Aggregating Capacity in FL through Successive Layer Training for Computationally-Constrained Devices

Kilian Pfeiffer, Ramin Khalili, Joerg Henkel

NeurIPS 2023 · Federated learning and privacy › Communication-efficient FL · approach: Efficient local training in FL · 3 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
use SLT instead of small model baselineFL training · ResNet20/CIFAR10 IID, s=0.125accuracy (%): 66.3 -> 74.1Improved
use SLT instead of FedRolexFL training · ResNet20/CIFAR10 IID, s=0.125accuracy (%): 21.7 -> 74.1Improved
use SLT instead of small model on easy datasetFL training · ResNet20/FEMNIST IID, s=0.125accuracy (%): 82.3 -> 84.4No clear effect
use small model instead of full modelFL training · CIFAR100 and TinyImageNet IIDaccuracy (%): up to 25.4 ppHurt
use SLT instead of small model at s=0.66FL training · DenseNet40/CIFAR100 IID, s=0.66accuracy (%): 55.9 -> 53.3Hurt
use SLT instead of small model under non-IIDFL training · ResNet20/CIFAR10 non-IID, s=0.125accuracy (%): 44.7 -> 52.4Improved
use SLT instead of small modelFL training · ResNet44/TinyImageNet IID, s=0.125accuracy (%): 19.8 -> 33.5Improved
use SLT instead of best baseline under heterogeneous constraintsFL training · ResNet44/TinyImageNet heterogeneous [0.125, 0.25, 0.5, 1.0]accuracy (%): 27.5 -> 35.9Improved
add a fourth constraint level for HeteroFLHeteroFL · ResNet44/TinyImageNet heterogeneousaccuracy (%): 24.1 -> 23.3Hurt
add a third constraint level (1.0) for FjORDFjORD · DenseNet40/CIFAR100 heterogeneousaccuracy (%): 38.7 -> 36.9Hurt
use SLT instead of small model on ImageNet 64x64FL training · ResNet56/ImageNet IID, s=0.125accuracy (%): 8.9 -> 24.2Improved
use fewer random subsets in FD (limit randomness) (ablation)FD · ResNet20/CIFAR10, s=0.25accuracy (%)Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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