Research map / Continual learning / Class-incremental
Incremental Few-Shot Learning with Attention Attractor Networks
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 |
|---|---|---|---|
| add static attractor regularizer (ablation) | LR classifier · tiered-ImageNet 200+5-way 1-shot | accuracy: 48.84 -> 55.36 | Helped |
| add static attractor regularizer (ablation) | MLP classifier · tiered-ImageNet 200+5-way 1-shot | accuracy: 41.22 -> 56.16 | Helped |
| add static attractor regularizer (ablation) | MLP classifier · mini-ImageNet 64+5-way 1-shot | accuracy: 49.36 -> 54.46 | Helped |
| replace static with attention (dynamic) attractor (ablation) | LR classifier · mini-ImageNet 64+5-way 1-shot | accuracy: 53.63 -> 55.31 | Helped |
| replace static with attention (dynamic) attractor (ablation) | MLP classifier · tiered-ImageNet 200+5-way 1-shot | accuracy: 56.16 -> 56.11 | No clear effect |
| replace static with attention (dynamic) attractor (ablation) | LR classifier · tiered-ImageNet 200+5-way 5-shot | accuracy: 65.53 -> 65.58 | No clear effect |
| use MLP instead of LR episodic classifier (with attention attractor) (ablation) | Attractor network · mini-ImageNet 64+5-way 1-shot | accuracy: 55.31 -> 54.95 | Hurt |
| train regularizer with truncated BPTT instead of recurrent back-propagation, then solve to convergence at test time (ablation) | Attractor network · mini-ImageNet 64+5-way | accuracy | Hurt |
Closest papers by meaning
- MetaFSCIL: A Meta-Learning Approach for Few-Shot Class Incremental Learning (CVPR 2022 · similarity 0.88)
- Pseudo Informative Episode Construction for Few-Shot Class-Incremental Learning (AAAI 2025 · similarity 0.88)
- M2SD:Multiple Mixing Self-Distillation for Few-Shot Class-Incremental Learning (AAAI 2024 · similarity 0.87)
- Few-Shot Incremental Learning With Continually Evolved Classifiers (CVPR 2021 · similarity 0.87)
- Tripartite Weight-Space Ensemble for Few-Shot Class-Incremental Learning (CVPR 2025 · similarity 0.86)
- Adaptive Decision Boundary for Few-Shot Class-Incremental Learning (AAAI 2025 · similarity 0.85)
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.