AI Paper Atlas

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

Ripple Shapley: Data Influence Attribution in One Federated Training Run

Dewen Zeng, Wenlong Tian, Haozhao Wang, Jianfeng Lu, Weijun Xiao, Zhiyong Xu

AAAI 2026 · Federated learning and privacy › Communication-efficient FL · approach: Communication-efficient heterogeneous FL · 2 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 Ripple Shapley instead of AFedSV+ for contribution evaluationfederated learning attribution · MNIST and CIFAR-10, 100 roundscumulative computation time (s): 61437.57 -> 984.98Improved
use Ripple Shapley instead of S-FedAvgfederated learning attribution · MNIST and CIFAR-10, 100 roundscumulative computation time (s): 4531.91 -> 984.98Improved
add Ripple Shapley attribution on top of plain trainingFedAvg · MNIST and CIFAR-10, 100 roundscomputation time (s): 480.91 -> 984.98Hurt
use Ripple Shapley aggregation weighting instead of baseline aggregationfederated aggregation · open-set label noiseaccuracy: over 10% average accuracy improvementImproved
use Ripple Shapley weighting on other poisoning settingsfederated aggregation · long-tailed, data noise (MNIST, CIFAR-10)accuracyNo clear effect
increase ripple propagation depth beyond 20 (ablation)Ripple Shapley · 100-round trajectories of representative samplesripple valueNo clear effect
add ripple term and low-rank Jacobian propagation to the drop termRipple Shapley · dynamic participation in FLattribution validityImproved

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