Research map / Federated learning and privacy / Personalized and heterogeneous FL
Federated Continuous Category Discovery and Learning
Lixu Wang, Chenxi Liu, Junfeng Guo, Qingqing Ye, Heng Huang, Haibo Hu, Wei Dong
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 semantic-weighted loss (SWL) aligning samples to global prototypes (ablation) | GPA · CIFAR-100 novel accuracy | novel accuracy (%): 37.3 -> 52.3 | Improved |
| add contrastive weighted loss (CWL) (ablation) | GPA · CIFAR-100 novel accuracy | novel accuracy (%): 42.9 -> 52.3 | Improved |
| add data mixup between confident and residual novel data (ablation) | GPA · CIFAR-100 novel accuracy | novel accuracy (%): 45.2 -> 52.3 | Improved |
| add model mixup instead of distillation (ablation) | GPA · CIFAR-100 known accuracy | known accuracy (%): 69.8 -> 71.9 | Improved |
| add prototype contrastive loss in known category learning (ablation) | GPA · CIFAR-100 novel accuracy | novel accuracy (%): 51.1 -> 52.3 | Improved |
| use PPM instead of MACC to estimate novel category number (ablation) | GPA · CIFAR-100 novel accuracy | novel accuracy (%): 39.8 -> 52.3 | Improved |
| use PPM instead of MACC or EMaCS for category number estimation | novel category number estimation · 27 category-count settings | average error rate (%): 48.9 -> 8.1 | Improved |
| apply GPA over Kmeans on FedProx, SCAFFOLD, Moon | FL algorithms · FC2DL one-stage | novel accuracy | Improved |
| use GPA instead of best baseline on CUB200 known categories | FC2DL · CUB200 known accuracy | known accuracy (%): 40.2 -> 40.3 | No clear effect |
Closest papers by meaning
- Federated Generalized Category Discovery (CVPR 2024 · similarity 0.88)
- FedSA: A Unified Representation Learning via Semantic Anchors for Prototype-based Federated Learning (AAAI 2025 · similarity 0.87)
- Federated Graph Learning under Domain Shift with Generalizable Prototypes (AAAI 2024 · similarity 0.87)
- Taming Cross-Domain Representation Variance in Federated Prototype Learning with Heterogeneous Data Domains (NeurIPS 2024 · similarity 0.86)
- Beyond Federated Prototype Learning: Learnable Semantic Anchors with Hyperspherical Contrast for Domain-Skewed Data (AAAI 2025 · similarity 0.86)
- Prior-Constrained Association Learning for Fine-Grained Generalized Category Discovery (AAAI 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.
