Research map / Federated learning and privacy / Personalized and heterogeneous FL
Decentralised Learning from Independent Multi-Domain Labels for Person Re-Identification
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Decouple client classification heads from aggregation and aggregate feature embeddings | FedReID · VIPeR and iLIDS | R1: 41.0 -> 46.2; 65.2 -> 69.7 | Helped |
| Use federated client-server updates instead of parameter-averaging trained local models | FedReID · VIPeR and iLIDS | R1: 19.9 -> 46.2; 41.8 -> 69.7 | Helped |
| Add local expert regularisation (ablation) | FedReID local client · Re-ID generalisation | R1 | Helped |
| Increase the number of collaborating clients to include multiple domains (ablation) | FedReID · Re-ID generalisation | R1 | Helped |
| Aggregate multiple clients rather than only one client (S=0.25) (ablation) | FedReID · Re-ID generalisation | R1 | Helped |
| Aggregate all clients rather than randomly selected clients (S=0.5) (ablation) | FedReID · Re-ID generalisation | R1 | Helped |
| Increase the number of local client training steps (ablation) | FedReID · Re-ID generalisation | R1 | Hurt |
| Increase noise scale beta for privacy protection (ablation) | FedReID aggregation · VIPeR | R1 | Hurt |
| Apply privacy noise in both aggregation and client-server updates rather than aggregation only (ablation) | FedReID · VIPeR | R1 | Hurt |
| Use supervised local data with a global model as the distillation expert | FedReID (direct) · Market-1501 and DukeMTMC-ReID | R1: 80.2 -> 90.3; 68.0 -> 77.4 | Helped |
| Use supervised local data with a global model as the distillation expert | FedReID (direct) · DukeMTMC-ReID | R1: 77.3 -> 77.4 | No clear effect |
| Train with federated updates rather than centralised joint training | FedReID · CIFAR-10 | Top-1 accuracy: 93.58$\pm$0.14 -> 93.31$\pm$0.11 | No clear effect |
Closest papers by meaning
- FedARKS: Federated Aggregation via Robust and Discriminative Knowledge Selection and Integration for Person Re-identification (AAAI 2026 · similarity 0.90)
- Generalising without Forgetting for Lifelong Person Re-Identification (AAAI 2021 · similarity 0.88)
- Domain-Aware Suppression and Aggregation for Federated DG ReID (AAAI 2026 · similarity 0.88)
- Diversity-Authenticity Co-constrained Stylization for Federated Domain Generalization in Person Re-identification (AAAI 2024 · similarity 0.87)
- From Laboratory to Real World: A New Benchmark Towards Privacy-Preserved Visible-Infrared Person Re-Identification (CVPR 2025 · similarity 0.86)
- Person De-reidentification: A Variation-guided Identity Shift Modeling (CVPR 2025 · 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.