Research map / Continual learning / Class-incremental
Forward Compatible Few-Shot Class-Incremental Learning
Check 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 |
|---|---|---|---|
| Train with virtual-prototype loss L2 in addition to L1 (ablation) | FACT · CIFAR100 FSCIL | accuracy | Helped |
| Add the forecasting loss L3 (ablation) | FACT with L1 and L2 · CIFAR100 FSCIL | accuracy | Helped |
| Add the symmetric regularization loss L4 (ablation) | FACT with L1, L2, and L3 · CIFAR100 FSCIL | accuracy | Helped |
| Use virtual prototypes during inference instead of ProtoNet inference (ablation) | FACT · CIFAR100 FSCIL | accuracy | Helped |
| Use L1 alone, without the virtual-prototype, forecasting, and symmetric losses (ablation) | FACT · CIFAR100 FSCIL | accuracy | Hurt |
| Increase the number of incremental examples per class (ablation) | FACT · miniImageNet FSCIL | last-session accuracy | Helped |
| Use manifold mixup, vanilla mixup, or CutMix for instance mixing (ablation) | FACT · CIFAR100 FSCIL | last-session accuracy | No clear effect |
| Increase the incremental shot count beyond 20 (ablation) | FACT · miniImageNet FSCIL | last-session accuracy | No clear effect |
| Increase virtual prototype count above 40 (ablation) | FACT · CIFAR100 FSCIL | final accuracy | No clear effect |
| Use a small trade-off parameter gamma (ablation) | FACT · CIFAR100 FSCIL | final accuracy | Helped |
Closest papers by meaning
- Tripartite Weight-Space Ensemble for Few-Shot Class-Incremental Learning (CVPR 2025 · similarity 0.89)
- Pseudo Informative Episode Construction for Few-Shot Class-Incremental Learning (AAAI 2025 · similarity 0.88)
- Adaptive Decision Boundary for Few-Shot Class-Incremental Learning (AAAI 2025 · similarity 0.88)
- Pre-trained Vision and Language Transformers Are Few-Shot Incremental Learners (CVPR 2024 · similarity 0.88)
- M2SD:Multiple Mixing Self-Distillation for Few-Shot Class-Incremental Learning (AAAI 2024 · similarity 0.87)
- Compositional Few-Shot Class-Incremental Learning (ICML 2024 · 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.