Research map / Continual learning / Online and replay-based
Isolation and Impartial Aggregation: A Paradigm of Incremental Learning without Interference
Read the paperCheck 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 stage-isolated classifiers with energy self-normalization (ablation) | ESN · Split CIFAR-100 | FAA: 80.21 -> 86.34 | Helped |
| Use stage-isolated classifiers with energy self-normalization (ablation) | ESN · Split CIFAR-100 | FF: 9.35 -> 4.76 | Helped |
| Add temperature calibration and voting (ablation) | ESN · Split CIFAR-100 | FAA: 85.73 -> 86.34 | Helped |
| Add temperature calibration and voting (ablation) | ESN · Split CIFAR-100 | FF: 4.88 -> 4.76 | Helped |
| Use separate stage-specific classifiers rather than sharing CAB (ablation) | ESN · Split CIFAR-100 | FAA: 83.94 -> 86.34 | Helped |
| Use separate stage-specific classifiers rather than sharing CAB (ablation) | ESN · Split CIFAR-100 | FF: 6.42 -> 4.76 | Helped |
| Vary the energy anchor across tested settings (ablation) | ESN · Split CIFAR-100 | FAA: 85.96, 85.60, 85.59, 86.20, 86.34, 86.25 | No clear effect |
| Replace CAB with VPT (ablation) | ESN · Split CIFAR-100 | FAA: 86.34 -> 85.55 | Hurt |
| Replace CAB with DER-ViT (ablation) | ESN · Split CIFAR-100 | FAA: 86.34 -> 83.43 | Hurt |
| Replace CAB with DER-ResNet50 (ablation) | ESN · Split CIFAR-100 | FAA: 86.34 -> 80.37 | Hurt |
| Use ESN instead of L2P | incremental learning with ViT-B/16 · CORe50 | FAA: 78.33 -> 91.80 | Helped |
| Use ESN instead of L2P | cross-domain class-incremental learning · Split DomainNet | FAA: 45.65 -> 68.76 | Helped |
Closest papers by meaning
- CAPrompt: Cyclic Prompt Aggregation for Pre-Trained Model Based Class Incremental Learning (AAAI 2025 · similarity 0.85)
- Tripartite Weight-Space Ensemble for Few-Shot Class-Incremental Learning (CVPR 2025 · similarity 0.85)
- MIND: Multi-Task Incremental Network Distillation (AAAI 2024 · similarity 0.84)
- Long-Tail Class Incremental Learning via Independent Sub-prototype Construction (CVPR 2024 · similarity 0.84)
- Adapter Merging with Centroid Prototype Mapping for Scalable Class-Incremental Learning (CVPR 2025 · similarity 0.84)
- Gradient Reweighting: Towards Imbalanced Class-Incremental Learning (CVPR 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.