Few-Shot Class-Incremental Learning
Abstract
The ability to incrementally learn new classes is crucial to the development of real-world artificial intelligence systems. In this paper, we focus on a challenging but practical few-shot class-incremental learning (FSCIL) problem. FSCIL requires CNN models to incrementally learn new classes from very few labelled samples, without forgetting the previously learned ones. To address this problem, we represent the knowledge using a neural gas (NG) network, which can learn and preserve the topology of the feature manifold formed by different classes. On this basis, we propose the TOpology-Preserving knowledge InCrementer (TOPIC) framework. TOPIC mitigates the forgetting of the old classes by stabilizing NG's topology and improves the representation learning for few-shot new classes by growing and adapting NG to new training samples. Comprehensive experimental results demonstrate that our proposed method significantly outperforms other state-of-the-art class-incremental learning methods on CIFAR100, miniImageNet, and CUB200 datasets.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
FSCIL demands that the model retains previously acquired knowledge while continually incorporating new classes, all while dealing with constraints of limited annotated samples for each class.
“This learning paradigm demands that the model retains previously acquired knowledge while continually incorporating new classes, all while dealing with the constraints of limited annotated samples for each class [12, 13, 14].”
From Few-Shot Class-Incremental Learning for Classification and Object Detection: A Survey (2025) · cites this paper as [13]
In FSCIL, the prevalent data partitioning method by Tao et al. divides mini ImageNet's 100 classes into 60 base and 40 incremental classes, with incremental classes segmented into 8 sessions, each with 5 classes and 5 training samples per class, forming a 5-way 5-shot setup.
“In FSCIL, the prevalent data partitioning method by Tao et al. [13] divides these 100 classes into 60 base and 40 incremental classes. These incremental classes are further segmented into 8 sessions, each with 5 classes and 5 training samples per class, forming a 5 -way 5 -shot setup.”
From Few-Shot Class-Incremental Learning for Classification and Object Detection: A Survey (2025) · cites this paper as [13]
For FSCIL, the common data partitioning approach by Tao et al. divides CIFAR-100's 100 classes into 60 base classes and 40 incremental classes, with incremental classes split into 8 sessions, each with 5 classes and 5 training samples per class, establishing a 5-way 5-shot format.
“For FSCIL, the common data partitioning approach by Tao et al. [13] divides these 100 classes into 60 base classes and 40 incremental classes. These incremental classes are further split into 8 sessions, each with 5 classes. Every class in these sessions has 5 training samples, establishing a 5 -way 5 -shot format.”
From Few-Shot Class-Incremental Learning for Classification and Object Detection: A Survey (2025) · cites this paper as [13]
Real-world applications often require models to learn new classes from limited samples while maintaining performance on base classes, which often represent high-frequency classes in the real world.
“However, real-world applications often require models to learn new classes from limited samples while maintaining performance on base classes, which often represent high-frequency classes in the real world [13, 54].”
From Few-Shot Class-Incremental Learning for Classification and Object Detection: A Survey (2025) · cites this paper as [13]
Cites
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.90 · Cited by 235,474 · see citing works
ImageNet classification with deep convolutional neural networks
Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2017) · Communications of the ACM · doi:10.1145/3065386 · Cited by 112,155 · see citing works
Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
Ren, S., He, K., Girshick, R., & Sun, J. (2017) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2016.2577031 · Cited by 56,035 · see citing works
Overcoming catastrophic forgetting in neural networks
Kirkpatrick, J., Pascanu, R., Rabinowitz, N., Veness, J., Desjardins, G., Rusu, A. A., Milan, K., Quan, J., Ramalho, T., Grabska-Barwinska, A., et al. (2017) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.1611835114 · Cited by 7,899 · see citing works
iCaRL: Incremental Classifier and Representation Learning
Rebuffi, S.-A., Kolesnikov, A., Sperl, G., & Lampert, C. H. (2017) · 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2017.587 · Cited by 3,988 · see citing works
Memory Aware Synapses: Learning What (not) to Forget
Aljundi, R., Babiloni, F., Elhoseiny, M., Rohrbach, M., & Tuytelaars, T. (2018) · Lecture Notes in Computer Science · doi:10.1007/978-3-030-01219-9_9 · Cited by 1,590 · see citing works
Competitive Hebbian Learning Rule Forms Perfectly Topology Preserving Maps
Martinetz, T. (1993) · ICANN ’93 · doi:10.1007/978-1-4471-2063-6_104 · Cited by 256 · see citing works
Flagged first, then by citations
Cited by
Sorted by citations
A Comprehensive Survey of Continual Learning: Theory, Method and Application
Wang, L., Zhang, X., Su, H., & Zhu, J. (2024) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2024.3367329 · Cited by 1,113 · see citing works
Topology-Preserving Class-Incremental Learning
Tao, X., Chang, X., Hong, X., Wei, X., & Gong, Y. (2020) · Lecture Notes in Computer Science · doi:10.1007/978-3-030-58529-7_16 · Cited by 162 · see citing works
Self-citationFew-Shot Class-Incremental Learning for Classification and Object Detection: A Survey
Zhang, J., Liu, L., Silvén, O., Pietikäinen, M., & Hu, D. (2025) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2025.3529038 · Cited by 84 · see citing works
“FSCIL was first carried out by TOPIC [13].”
InfluentialContinual Learning for Generative Retrieval over Dynamic Corpora
Chen, J., Zhang, R., Guo, J., de Rijke, M., Chen, W., Fan, Y., & Cheng, X. (2023) · Proceedings of the 32nd ACM International Conference on Information and Knowledge Management · doi:10.1145/3583780.3614821 · Cited by 40 · see citing works
Few-Shot Class-Incremental Learning for Retinal Disease Recognition
Zhang, J., Zhao, P., Zhao, Y., Li, C., & Hu, D. (2025) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2024.3457915 · Cited by 23 · see citing works
FLAR: A Unified Prototype Framework for Few-sample Lifelong Active Recognition
Fan, L., Xiong, P., Wei, W., & Wu, Y. (2021) · 2021 IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv48922.2021.01511 · Cited by 6 · see citing works
A Comprehensive Study of Class Incremental Learning Algorithms for Visual Tasks
Belouadah, E., Popescu, A., & Kanellos, I. (2020) · arXiv · doi:10.48550/arxiv.2011.01844 · Cited by 235 · see citing works
- Authors
- Tao, X., Hong, X., Chang, X., Dong, S., Wei, X., & Gong, Y.
- Affiliations
- Xi'an Jiaotong University (CN)
- Peng Cheng Laboratory (CN)
- Venue
- 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
- Publisher
- IEEE
- Published
- 2020-06
- Pages
- 12180-12189
- ISSN
- 1063-6919 (ISSN-L)
- OpenAlex
- W3035342403
- Cited by
- 499
- References
- 7
- Citations by year
- 2026: 51 · 2025: 122 · 2024: 114 · 2023: 109 · 2022: 55 · 2021: 41 · 2020: 7
- FWCI
- 20.25
- Type
- conference paper
- Language
- en
- Open access
- closed
- Licence
- Locations
- IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings (publishedVersion)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Domain Adaptation and Few-Shot Learning
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Physical Sciences › Computer Science › Artificial Intelligence › Machine Learning and Data Classification
- Keywords
- few-shot class-incremental learning, catastrophic forgetting, neural gas network, topology preservation, feature manifold, representation learning, continual learning, convolutional neural networks
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
