Research map / Federated learning and privacy / Personalized and heterogeneous FL
FedDAP: Domain-Aware Prototype Learning for Federated Learning under Domain Shift
Huy Q. Le, Loc X. Nguyen, Yu Qiao, Seong Tae Kim, Eui-Nam Huh, Choong Seon Hong
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 cross-domain contrastive loss L_CPCL on top of DPA (ablation) | FedDAP · DomainNet | average accuracy (%): 62.61 -> 65.20 | Improved |
| add cross-domain contrastive loss L_CPCL on top of DPA (ablation) | FedDAP · Office-10 | average accuracy (%): 68.53 -> 72.53 | Improved |
| add cross-domain contrastive loss L_CPCL on top of DPA (ablation) | FedDAP · PACS | average accuracy (%): 78.74 -> 84.63 | Improved |
| add domain-consistent alignment L_DPA on top of CPCL (ablation) | FedDAP · DomainNet | average accuracy (%): 62.86 -> 65.20 | Improved |
| add domain-consistent alignment L_DPA on top of CPCL (ablation) | FedDAP · Office-10 | average accuracy (%): 62.18 -> 72.53 | Improved |
| add domain-consistent alignment L_DPA on top of CPCL (ablation) | FedDAP · PACS | average accuracy (%): 81.87 -> 84.63 | Improved |
| use similarity-weighted fusion instead of simple averaging for domain prototypes (ablation) | FedDAP · DomainNet | average accuracy (%): 64.15 -> 65.20 | Improved |
| use similarity-weighted fusion instead of simple averaging for domain prototypes (ablation) | FedDAP · PACS | average accuracy (%): 83.22 -> 84.63 | Improved |
| use similarity-weighted fusion instead of simple averaging for domain prototypes (ablation) | FedDAP · Office-10 | average accuracy (%): 71.49 -> 72.53 | Improved |
| add Gaussian noise (differential privacy) to local prototypes (ablation) | FedDAP · DomainNet | average accuracy (%): 65.20 -> 64.38 | Hurt |
| add Gaussian noise (differential privacy) to local prototypes (ablation) | FedDAP · PACS | average accuracy (%): 84.63 -> 83.63 | Hurt |
| use domain-specific prototypes with dual alignment vs FedAvg | FedDAP · DomainNet (leave-one-domain-out) | average accuracy (%): 43.90 -> 49.52 | Improved |
Closest papers by meaning
- Taming Cross-Domain Representation Variance in Federated Prototype Learning with Heterogeneous Data Domains (NeurIPS 2024 · similarity 0.92)
- FedSA: A Unified Representation Learning via Semantic Anchors for Prototype-based Federated Learning (AAAI 2025 · similarity 0.90)
- Beyond Federated Prototype Learning: Learnable Semantic Anchors with Hyperspherical Contrast for Domain-Skewed Data (AAAI 2025 · similarity 0.90)
- Federated Graph Learning under Domain Shift with Generalizable Prototypes (AAAI 2024 · similarity 0.89)
- FedWon: Triumphing Multi-domain Federated Learning Without Normalization (ICLR 2024 · similarity 0.87)
- Federated Domain Generalization with Data-free On-server Matching Gradient (ICLR 2025 · similarity 0.87)
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.
