Research map / Continual learning / Class-incremental
Large Scale 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 |
|---|---|---|---|
| Add the two-parameter bias correction layer to the distillation-based incremental classifier (ablation) | CIFAR-100 incremental classifier · CIFAR-100 (100 classes, 20-class increments) | accuracy (%): 40.34 -> 56.69 | Helped |
| Add distillation loss to classification loss (ablation) | CIFAR-100 baseline-1 · CIFAR-100 (100 classes, 20-class increments) | accuracy (%): 39.83 -> 40.34 | Helped |
| Use an 8:2 rather than 9:1 train/validation split for old exemplars (ablation) | BiC · CIFAR-100 (100 classes, 20-class increments) | accuracy (%): 56.69 -> 54.31 | Hurt |
| Use a 7:3 rather than 9:1 train/validation split for old exemplars (ablation) | BiC · CIFAR-100 (100 classes, 20-class increments) | accuracy (%): 56.69 -> 53.74 | Hurt |
| Use a 6:4 rather than 9:1 train/validation split for old exemplars (ablation) | BiC · CIFAR-100 (100 classes, 20-class increments) | accuracy (%): 56.69 -> 51.17 | Hurt |
| Select stored exemplars randomly instead of using iCaRL selection (ablation) | BiC · CIFAR-100 (100 classes, 20-class increments) | accuracy (%): 56.69 -> 55.55 | Hurt |
| Use BiC instead of EEIL | Incremental classifier · ImageNet-1000 | accuracy (%): 73.2 -> 52.0 | Helped |
| Use BiC instead of iCaRL | Incremental classifier · ImageNet-1000 | accuracy (%): 73.2 -> 44.0 | Helped |
| Use BiC instead of iCaRL | Incremental classifier · Celeb-10000 | accuracy (%): 87.98 -> 65.56 | Helped |
| Use BiC instead of EEIL | Incremental classifier · CIFAR-100, 5- and 10-class increments | accuracy | Mixed |
Closest papers by meaning
- Learning a Unified Classifier Incrementally via Rebalancing (CVPR 2019 · similarity 0.92)
- Gradient Reweighting: Towards Imbalanced Class-Incremental Learning (CVPR 2024 · similarity 0.92)
- Long-Tail Class Incremental Learning via Independent Sub-prototype Construction (CVPR 2024 · similarity 0.88)
- Defying Imbalanced Forgetting in Class Incremental Learning (AAAI 2024 · similarity 0.88)
- Class Incremental Learning with Multi-Teacher Distillation (CVPR 2024 · similarity 0.87)
- Addressing Imbalanced Domain-Incremental Learning through Dual-Balance Collaborative Experts (ICML 2025 · 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.