Research map / Continual learning / Class-incremental
Few-Shot Class-Incremental Learning via Training-Free Prototype Calibration
Qi-Wei Wang, Da-Wei Zhou, Yi-Kai Zhang, De-Chuan Zhan, Han-Jia Ye
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 |
|---|---|---|---|
| Fuse new-class prototypes with similarity-weighted base prototypes | Prototype-based FSCIL · miniImageNet session 1 | NAcc: 19.60 -> 38.00 | Improved |
| Fuse new-class prototypes with similarity-weighted base prototypes | Prototype-based FSCIL · miniImageNet session 1 | HMean: 30.72 -> 50.04 | Improved |
| Use TEEN prototype calibration | Few-shot classification · miniImageNet 5-way 1-shot | accuracy: 68.57 -> 65.70 | Hurt |
| Use TEEN prototype calibration | Few-shot classification · miniImageNet 5-way 5-shot | accuracy: 82.88 -> 83.11 | Improved |
| Use TEEN prototype calibration | Few-shot classification · CUB 5-way 1-shot | accuracy: 79.56 -> 81.44 | Improved |
| Use TEEN prototype calibration | Few-shot classification · CUB 5-way 5-shot | accuracy: 90.67 -> 91.04 | Improved |
| Remove similarity-based softmax weights and use unweighted top-K base prototypes (SimTEEN) (ablation) | TEEN · FSCIL benchmarks | new-class accuracy | Hurt |
| Vary calibration coefficient alpha (ablation) | TEEN · FSCIL benchmarks | new-class accuracy | Mixed |
| Use a relatively smaller similarity temperature tau to smooth the weights (ablation) | TEEN · FSCIL benchmarks | average accuracy | Improved |
| Add TEEN calibration to FACT (ablation) | FACT · FSCIL benchmarks | average accuracy | No clear effect |
| Add TEEN calibration to FACT (ablation) | FACT · FSCIL benchmarks | new-class accuracy | Improved |
Closest papers by meaning
- Few-Shot Class-Incremental Learning by Sampling Multi-Phase Tasks (TPAMI 2022 · similarity 0.94)
- Few-Shot Incremental Learning With Continually Evolved Classifiers (CVPR 2021 · similarity 0.92)
- Few-Shot Class-Incremental Learning via Class-Aware Bilateral Distillation (CVPR 2023 · similarity 0.92)
- Tripartite Weight-Space Ensemble for Few-Shot Class-Incremental Learning (CVPR 2025 · similarity 0.90)
- Adaptive Decision Boundary for Few-Shot Class-Incremental Learning (AAAI 2025 · similarity 0.90)
- Enhancing Few-Shot Class-Incremental Learning via Training-Free Bi-Level Modality Calibration (CVPR 2025 · similarity 0.89)
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.
