Research map / Continual learning / Class-incremental
Few-shot class-incremental prototypes
Adds new classes from few samples using prototype calibration, meta-learning and orthogonal features.
47 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2019–2026) · an approach in Class-incremental, keyword cluster “shot · incremental cil · feature”
State History Benchmarks Tried Checklist Preprints Who Export
Where it stands
Still moving The best reported result is still improving on 1 of 1 tracked benchmark.
13 papers in 2023–24 and 17 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 14.1% to 12.6%.
Open questions the data points to
- CUB200 session 10 (accuracy): best went from 62.5 (up to 2024) to 84.8 in 2025 — SEC-Prompt:SEmantic Complementary Prompting for Few-Shot Class-Incremental Learning (CVPR 2025 · 5 citations)
How this is decided
A benchmark is tracked when at least four papers of this approach report their own method on the same dataset and metric in at least two different years. Its record is the best number reported up to 2024 compared with the best reported in 2025–26: still improving means at least 1% better, flat means at most 0.2% better with at least two papers from 2025–26 trying. Numbers on a different scale from the rest (a fraction among percentages) and far outliers are left out. Results come from different papers with different setups, so they are signals, not head-to-head comparisons. The share of papers is counted at ICML, NeurIPS, ICLR, CVPR and AAAI.
How it got here
Most cited paper of each year
- 2019 · Incremental Few-Shot Learning with Attention Attractor Networks (NeurIPS 2019 · 30 citations)
- 2020 · Few-Shot Class-Incremental Learning (CVPR 2020 · 470 citations)
- 2021 · Few-Shot Incremental Learning With Continually Evolved Classifiers (CVPR 2021 · 363 citations)
- 2022 · Forward Compatible Few-Shot Class-Incremental Learning (CVPR 2022 · 282 citations)
- 2023 · Learning With Fantasy: Semantic-Aware Virtual Contrastive Constraint for Few-Shot Class-Incremental Learning (CVPR 2023 · 113 citations)
- 2024 · OrCo: Towards Better Generalization via Orthogonality and Contrast for Few-Shot Class-Incremental Learning (CVPR 2024 · 45 citations)
- 2025 · Adaptive Decision Boundary for Few-Shot Class-Incremental Learning (AAAI 2025 · 6 citations)
- 2026 · SAFA-SNN: Sparsity-Aware On-Device Few-Shot Class-Incremental Learning with Fast-Adaptive Structure of Spiking Neural Network (ICLR 2026 · 0 citations)
Earliest papers here
- Incremental Few-Shot Learning with Attention Attractor Networks (NeurIPS 2019 · 30 citations)
- Few-Shot Class-Incremental Learning (CVPR 2020 · 470 citations)
- Few-Shot Incremental Learning With Continually Evolved Classifiers (CVPR 2021 · 363 citations)
Benchmark records
The best number each year that beat every earlier one, among papers of this approach reporting their own method. Numbers are extracted by AI from each paper's text; check the paper before citing.
CUB200 session 10 · accuracy (higher is better)
5 papers, 2021–2025 · still improving
| Year | Best so far | Set by |
|---|---|---|
| 2021 | 52.28 | Few-Shot Incremental Learning With Continually Evolved Classifiers (CVPR 2021 · 363 citations) |
| 2023 | 62.5 | Learning With Fantasy: Semantic-Aware Virtual Contrastive Constraint for Few-Shot Class-Incremental Learning (CVPR 2023 · 113 citations) |
| 2025 | 84.8 | SEC-Prompt:SEmantic Complementary Prompting for Few-Shot Class-Incremental Learning (CVPR 2025 · 5 citations) |
What has been tried
35 papers here report 389 measured changes (a component added, removed or swapped, compared with a baseline): helped 76%, no clear effect 4%, hurt 16%, mixed 4%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| Add ORTH to PSCL + CE | mini-ImageNet · aHM | 36.17 -> 58.12 | Helped | OrCo: Towards Better Generalization via Orthogonality and Contrast for Few-Shot Class-Incremental Learning (CVPR 2024 · 45 citations) |
| Add base-session pretraining before the self-distillation stage | CUB200 session 10 · accuracy | 52.37 -> 60.96 | Helped | M2SD:Multiple Mixing Self-Distillation for Few-Shot Class-Incremental Learning (AAAI 2024 · 14 citations) |
| Add adaptive decision boundary without inter-class constraint | miniImageNet · session 8 accuracy | 50.81 -> 54.67 | Helped | Adaptive Decision Boundary for Few-Shot Class-Incremental Learning (AAAI 2025 · 6 citations) |
| add novel-novel diversity criteria for selecting pseudo novel classes | CUB200 · performance dropping (PD) | 16.41% vs. 17.43% | Helped | Pseudo Informative Episode Construction for Few-Shot Class-Incremental Learning (AAAI 2025 · 5 citations) |
| Replace Coda-Prompt with D-Prompt and Adaptive Hierarchical Query | ImageNet-R session 10 · accuracy | 66.4 -> 73.1 | Helped | SEC-Prompt:SEmantic Complementary Prompting for Few-Shot Class-Incremental Learning (CVPR 2025 · 5 citations) |
| Add MeCs for prototype construction | CoraFull, CS, and Computers · accuracy | Helped | An Efficient Memory Module for Graph Few-Shot Class-Incremental Learning (NeurIPS 2024 · 1 citation) | |
| Vary base-session anchor selection among closest-to-prototype, k-means, and random-closest strategies instead of random sampling | BCI-IV 2a, Session 2 · Macro-F1 (All Classes) | 37.04±1.24; 38.65 ± 1.26; 36.97±1.28; 38.36±1.34; 38.55±1.30; 38.05±1.61 | Mixed | AnchorInv: Few-Shot Class-Incremental Learning of Physiological Signals via Feature Space-Guided Inversion (AAAI 2025 · 0 citations) |
| Add inference metric alteration (IM) | CIFAR100, session 8 · incremental class accuracy (%) | 15.45 -> 22.18 | Helped | Attraction Diminishing and Distributing for Few-Shot Class-Incremental Learning (CVPR 2025 · 0 citations) |
| Use Tri-WE with φ0, φold, and φall instead of no weight-space ensemble | miniImageNet · average accuracy (%) | 67.92 -> 70.62 | Helped | Tripartite Weight-Space Ensemble for Few-Shot Class-Incremental Learning (CVPR 2025 · 0 citations) |
| add PAMMD reward | CIFAR-100 · last-session accuracy improvement (%) | +1.24 | Helped | Diffusion-Classifier Synergy: Reward-Aligned Learning via Mutual Boosting Loop for FSCIL (NeurIPS 2025 · 0 citations) |
| add the composition module | CUB200 last incremental session · Overall accuracy | 57.18 -> 59.25 | Helped | Compositional Few-Shot Class-Incremental Learning (ICML 2024 · 0 citations) |
What reviewers will expect
Datasets and baselines named in the 9 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- CIFAR100 · 6 of 9 papers
- miniImageNet · 5 of 9 papers
- CUB200 · 5 of 9 papers
Baselines
- TEEN · 5 of 9 papers
- TOPIC · 4 of 9 papers
- CEC · 4 of 9 papers
- FACT · 4 of 9 papers
- NC-FSCIL · 4 of 9 papers
- SAVC · 4 of 9 papers
- SoftNet · 3 of 9 papers
- DyCR · 3 of 9 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 30 papers; from OpenAlex and the papers' affiliations).
- South China University of Technology · 2 papers
- Nanyang Technological University · 2 papers
- University of Oxford · 2 papers
- Peng Cheng Laboratory · 2 papers
- Max Planck Institute for Informatics · 2 papers
- Peking University · 2 papers
- Korea Advanced Institute of Science and Technology · 2 papers
Use it
Download BibTeX of 39 key papersOpen the topic in the interactive map
Get alerts. Describe your idea in scime atlas and save it (free scime account): you get a weekly email when new preprints land close to it.
Related approaches
Same keyword cluster
- Exemplar-efficient class-incremental learning (111 papers)
- Class-balanced knowledge retention (73 papers)
- Incremental object detection with replay (32 papers)
Elsewhere in the topic
- Unknown-class mining for incremental segmentation (23 papers)
- Distillation-based continual segmentation (17 papers)
Benchmark numbers, ablations and baseline numbers are extracted by AI from each paper's full text and can be wrong; the paper is the authority. Data: results index of 2026-11-02, citation counts from OpenAlex (2026-10-09).