Research map / Continual learning / Class-incremental
Learning a Unified Classifier Incrementally via Rebalancing
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 |
|---|---|---|---|
| Use Ours-CNN instead of iCaRL-NME | Unified classifier incremental learning · CIFAR100 (1 phase) | top-1 accuracy: 62.34% vs. 59.13% | Helped |
| Add cosine normalization (ablation) | Proposed incremental classifier · CIFAR100 (5 phases) | classification accuracy | Helped |
| Add the less-forget constraint to cosine normalization (ablation) | Proposed incremental classifier · CIFAR100 (5 phases) | classification accuracy | Helped |
| Add inter-class separation (ablation) | Proposed incremental classifier · CIFAR100 (5 phases) | classification accuracy | Helped |
| Add class-balance finetuning on reserved samples (ablation) | Proposed incremental classifier · CIFAR100 (5 phases) | classification accuracy | Helped |
| Increase the number of reserved samples per class (ablation) | Proposed method and iCaRL · CIFAR100 (5 phases) | classification accuracy | Helped |
| Use adaptive rather than fixed distillation-loss weight (ablation) | Proposed incremental classifier · CIFAR100 (5 and 10 phases) | classification accuracy | Helped |
Closest papers by meaning
- Large Scale Incremental Learning (CVPR 2019 · similarity 0.92)
- Gradient Reweighting: Towards Imbalanced Class-Incremental Learning (CVPR 2024 · similarity 0.91)
- Class-Incremental Learning With Strong Pre-Trained Models (CVPR 2022 · similarity 0.89)
- Rebalancing Multi-Label Class-Incremental Learning (AAAI 2025 · similarity 0.87)
- Cross-Class Feature Augmentation for Class Incremental Learning (AAAI 2024 · similarity 0.86)
- Prediction Error-based Classification for Class-Incremental Learning (ICLR 2024 · 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.