Research map / Federated learning and privacy / Personalized and heterogeneous FL
FedARC: Anchor-Guided Residual Compensation for Data and Model Heterogeneous Federated Learning
Chentao Lu, Xuhao Ren, Dawei xu, Chuan Zhang, Liehuang Zhu
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add adaptive residual compensation on top of feature representation fusion (ablation) | FedARC · CIFAR-100 (HtFE8) | accuracy (%): 40.13 -> 43.64 (+3.51) | Improved |
| add semantic anchor alignment on top of feature representation fusion (ablation) | FedARC · CIFAR-100 (HtFE8) | accuracy (%): 40.13 -> 41.39 (+1.26) | Improved |
| combine residual compensation and semantic anchor alignment (ablation) | FedARC with fusion · CIFAR-100 (HtFE8) | accuracy (%): 40.13 -> 44.21 (+4.08) | Improved |
| vary anchor regularization lambda away from 1 (0.1, 5, 10) (ablation) | FedARC · CIFAR-100 (HtFE8) | accuracy (%): 44.21 -> 42.92 (lambda 0.1); 43.14 (lambda 5); 42.46 (lambda 10) | Hurt |
| increase momentum kappa from 0.1 to 0.5 or 1 (ablation) | FedARC · CIFAR-100 (HtFE8) | accuracy (%): 44.21 -> 41.87 (kappa 0.5); 40.46 (kappa 1) | Hurt |
| lower momentum kappa to 0.05 (ablation) | FedARC · CIFAR-100 (HtFE8) | accuracy (%): 44.21 -> 43.65 | Hurt |
| increase model heterogeneity from HtFE2 to HtFE9 | FedARC · CIFAR-100 cross-device | accuracy (%): 44.36 -> 37.28 | Hurt |
| increase local epochs E | mutual learning baselines (pFedES, FedKD) · CIFAR-100 cross-silo | test accuracy | Hurt |
| use FedARC over FedMRL/pFedES in communication cost | mutual learning HtFL · CIFAR-100 (HtFE8) | upload per round (MB): 50.75 -> 39.87 | Improved |
| increase feature dimension d1 from 64 to 256 (ablation) | most HtFL methods incl. FedARC · CIFAR-100 (HtFE8) | test accuracy | Improved |
| FedARC vs pFedES on heterogeneous full models HtM10 | FedARC · CIFAR-100 cross-device, HtM10 | test accuracy (%): 39.74 vs pFedES 40.04 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- pFedES — pFedES: Generalized Proxy Feature Extractor Sharing for Model Heterogeneous Personalized Federated Learning (AAAI 2025) · better on 0 of 1 shared result
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
- FedSA: A Unified Representation Learning via Semantic Anchors for Prototype-based Federated Learning (AAAI 2025 · similarity 0.88)
- Overcoming Data and Model heterogeneities in Decentralized Federated Learning via Synthetic Anchors (ICML 2024 · similarity 0.87)
- Federated Model Heterogeneous Matryoshka Representation Learning (NeurIPS 2024 · similarity 0.87)
- FedDAT: An Approach for Foundation Model Finetuning in Multi-Modal Heterogeneous Federated Learning (AAAI 2024 · similarity 0.87)
- An Upload-Efficient Scheme for Transferring Knowledge From a Server-Side Pre-trained Generator to Clients in Heterogeneous Federated Learning (CVPR 2024 · similarity 0.86)
- A Swiss Army Knife for Heterogeneous Federated Learning: Flexible Coupling via Trace Norm (NeurIPS 2024 · similarity 0.86)
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.
