Research map / Continual learning / Class-incremental
Cross-View Lewis Weight Fusion Empowering Exemplar Replay for Federated Class-Incremental Learning
Zhuang Qi, Ying-Peng Tang, Lei Meng, Xiaoxiao Li, Han Yu, Xiangxu Meng
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-view Lewis weight fusion (p=1) (ablation) | Re-Fed+ · CIFAR10 (5 tasks, 20 clients, beta=0.5) | Top-1 accuracy: 52.45 -> 56.11 | Improved |
| add cross-view Lewis weight fusion (p=1) (ablation) | FedCBDR · CIFAR10 (5 tasks, 20 clients, beta=0.5) | Top-1 accuracy: 56.94 -> 60.94 | Improved |
| add cross-view Lewis weight fusion (p=1) (ablation) | FedCBDR · CIFAR100 (5 tasks, 20 clients, beta=0.5) | Top-1 accuracy: 42.52 -> 43.21 | Improved |
| add frequency-aware weighted training on top of CV-LWF (p=1) (ablation) | Re-Fed+ · CIFAR10 (5 tasks, 20 clients, beta=0.5) | Top-1 accuracy: 56.11 -> 57.30 | Improved |
| add frequency-aware weighted training on top of CV-LWF (p=1) (ablation) | FedCBDR · CIFAR10 (5 tasks, 20 clients, beta=0.5) | Top-1 accuracy: 60.94 -> 62.28 | Improved |
| add frequency-aware weighted training on top of CV-LWF (p=1) (ablation) | FedCBDR · CIFAR100 (5 tasks, 20 clients, beta=0.5) | Top-1 accuracy: 43.21 -> 43.54 | Improved |
| use Lewis p=2 instead of p=1 (ablation) | Re-Fed+ + CV-LWF · CIFAR10 (5 tasks, 20 clients, beta=0.5) | Top-1 accuracy: 56.11 -> 56.40 | No clear effect |
| increase views from 2 to 5 (ablation) | CLIFR · CIFAR10 (20 clients) | Top-1 accuracy: 52.34 -> 54.97 | Improved |
| increase views from 5 to 10 (ablation) | CLIFR · CIFAR10 (20 clients) | Top-1 accuracy: 54.97 -> 51.66 | Hurt |
| increase views from 5 to 10 (ablation) | CLIFF · CIFAR10 (20 clients) | Top-1 accuracy: 56.63 -> 57.67 | Improved |
| increase views from 5 to 10 (ablation) | CLIFF · CIFAR100 (20 clients) | Top-1 accuracy: 44.20 -> 44.28 | No clear effect |
| increase number of views from 2 to 10 (selection time) (ablation) | CLIFR · CIFAR10 sample selection time (s) | time (s): 2.251 -> 6.825 | Hurt |
Closest papers by meaning
- Class-wise Balancing Data Replay for Federated Class-Incremental Learning (NeurIPS 2025 · similarity 0.88)
- Tripartite Weight-Space Ensemble for Few-Shot Class-Incremental Learning (CVPR 2025 · similarity 0.86)
- Federated Class-Incremental Learning: A Hybrid Approach Using Latent Exemplars and Data-Free Techniques to Address Local and Global Forgetting (ICLR 2025 · similarity 0.85)
- Autoencoder-Based Hybrid Replay for Class-Incremental Learning (ICML 2025 · similarity 0.85)
- Tensor Decomposition Based Memory-Efficient Incremental Learning (ICML 2025 · similarity 0.84)
- Elastic Feature Consolidation For Cold Start Exemplar-Free Incremental Learning (ICLR 2024 · similarity 0.84)
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.
