Research map / Continual learning / Class-incremental
Self-Organizing Pathway Expansion for Non-Exemplar Class-Incremental Learning
Kai Zhu, Kecheng Zheng, Ruili Feng, Deli Zhao, Yang Cao, Zheng-Jun Zha
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 class-specific pathway organization (CPO) in the base phase (ablation) | NECIL baseline · TinyImageNet (P=5) | average incremental accuracy: +3.04% | Improved |
| add class-specific pathway organization (CPO) in the base phase (ablation) | NECIL baseline · TinyImageNet (P=20) | average incremental accuracy: +8.31% | Improved |
| add pathway-guided feature update (PFU) (ablation) | SOPE with CPO · TinyImageNet | average incremental accuracy: about 2 points average | Improved |
| add incremental pathway update (IPU) (ablation) | SOPE with CPO · TinyImageNet | average incremental accuracy: about 1 point average | Improved |
| use soft learnable pathway threshold instead of hard-threshold pathway methods (Hat, Iap) or Piggy/RPS (ablation) | pathway strategies for NECIL · TinyImageNet | average incremental accuracy | Improved |
| divide channels into only 2 groups (ablation) | CPO · CIFAR-100 | incremental accuracy | Hurt |
| raise maximum sparse rate to a high value (ablation) | CPO · CIFAR-100 | incremental accuracy | Hurt |
| increase sparsity between 0.3 and 0.45 (ablation) | CPO · CIFAR-100 | incremental accuracy | Improved |
Closest papers by meaning
- Self-Sustaining Representation Expansion for Non-Exemplar Class-Incremental Learning (CVPR 2022 · similarity 0.89)
- Harnessing Neural Unit Dynamics for Effective and Scalable Class-Incremental Learning (ICML 2024 · similarity 0.86)
- Non-exemplar Online Class-Incremental Continual Learning via Dual-Prototype Self-Augment and Refinement (AAAI 2024 · similarity 0.85)
- Adaptive Discovering and Merging for Incremental Novel Class Discovery (AAAI 2024 · similarity 0.85)
- M2SD:Multiple Mixing Self-Distillation for Few-Shot Class-Incremental Learning (AAAI 2024 · similarity 0.84)
- Expandable Subspace Ensemble for Pre-Trained Model-Based 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.
