AI Paper Atlas

Research map / Continual learning / Online and replay-based

FedAGC: Federated Continual Learning with Asymmetric Gradient Correction

Chengchao Zhang, Fanhua Shang, Hongying Liu, Liang Wan, Wei Feng

ICCV 2025 · Continual learning › Online and replay-based · approach: Unified and unsupervised continual frameworks · 1 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
add asymmetric correction gradient updating (ACGU) (ablation)FedAGC · CIFAR-100 + 7 datasets, Non-IID (10 clients)Aavg: 65.61 -> 79.71Improved
add ACGU (ablation)FedAGC · Single CIFAR-100, IID (10 clients)Aavg: 63.53 -> 70.16Improved
add group aggregation (GAGU) (ablation)FedAGC · CIFAR-100 + 7 datasets, Non-IID (10 clients)Aavg: 76.95 -> 79.71Improved
add group aggregation (GAGU) (ablation)FedAGC · Single CIFAR-100, IID (10 clients)Aavg: 66.58 -> 70.16Improved
add centroid-based representative sample selection (CBRSS) (ablation)FedAGC · CIFAR-100 + 7 datasets, Non-IID (10 clients)Aavg: 78.56 -> 79.71Improved
add CBRSS (ablation)FedAGC · Single CIFAR-100, IID (10 clients)Aavg: 69.75 -> 70.16Improved
replace FedAvg with FedAGCFedAvg · CIFAR-100 + 7 datasets, Non-IID (10 clients)Aavg: 61.83 -> 79.71Improved
replace FedMes with FedAGCFedMes · CIFAR-100 + 7 datasets, Non-IID (10 clients)Aavg: 73.54 -> 79.71Improved
replace FedMes with FedAGCFedMes · Single CIFAR-100, IID (10 clients)Aavg: 67.13 -> 70.16Improved
add personalization penalty (FedProx)FedAvg · CIFAR-100 + 7 datasets, Non-IID (10 clients)Aavg: 61.83 -> 65.00Improved
add personalization penalty (FedProx)FedAvg · Single CIFAR-100, IID (10 clients)Aavg: 62.63 -> 62.97No clear effect
vary gamma and lambda away from 0.5 and 0.01 (ablation)FedAGC · Non-IID and IID CIFAR-100AavgHurt

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