Few-Shot Class-Incremental Learning Based on Feature Distribution Learning

doi:10.3778/j.issn.1002-8331.2205-0385

Abstract

Abstract: This paper focuses on a very challenging problem：few-shot class-incremental learning for deep neural networks, where the deep neural network model can gradually learn new knowledge from a small number of samples without forgetting the learned knowledge. To balance the model’s memory of old knowledge and learning of new knowledge, it proposes a few-shot class-incremental learning method based on feature distribution learning. First, it learns the model on the base classes to obtain a well-performing feature extractor and take the feature distribution information as the learned knowledge. Then, it maps the learned knowledge together with the features of the novel classes into a low-dimensional subspace to review old knowledge and learn new knowledge uniformly. Finally, within the subspace, it also generates classification weight initializations for each novel class to improve the adaptability of the model to novel classes. Extensive experiments show that the method can effectively alleviate the model’s forgetting of learned knowledge and improve the model’s adaptability to new knowledge.

Key words: few-shot class-incremental learning, deep neural network, incremental learning

摘要： 关注了一个非常具有挑战性的问题：深度神经网络的小样本类增量学习。其中深度神经网络模型可以从少量的样本中逐步学习新知识，同时不会忘记已学习的旧知识。为了平衡模型对旧知识的记忆和对新知识的学习，提出了一个基于特征分布学习的小样本类增量学习方法。在基类上学习模型以获得一个性能良好的特征提取器，并使用每类的特征分布信息来表示知识。将已学习的知识与新类的特征一起映射到一个新的低维子空间中，以统一地回顾旧知识与学习新知识。在子空间内，还为每个新类生成了分类权值初始化，以提高模型对新类的适应性。大量实验表明，该方法可以有效地减轻模型对已学习知识的遗忘，同时提高模型对新知识的适应性。

关键词: 小样本类增量学习, 深度神经网络, 增量学习

YAO Guangle, ZHU Juntao, ZHOU Wenlong, ZHANG Guiyu, ZHANG Wei, ZHANG Qian. Few-Shot Class-Incremental Learning Based on Feature Distribution Learning[J]. Computer Engineering and Applications, 2023, 59(14): 151-157.

姚光乐, 祝钧桃, 周文龙, 张贵宇, 张伟, 张谦. 基于特征分布学习的小样本类增量学习[J]. 计算机工程与应用, 2023, 59(14): 151-157.

References

[1] TAO X，HONG X，CHANG X，et al.Few-shot class-incremental learning[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2020：12180-12189.
[2] DOUILLARD A，CORD M，OLLION C，et al.PODNet：pooled outputs distillation for small-tasks incremental learning[C]//European Conference on Computer Vision，2020.
[3] GIDARIS S，KOMODAKIS N.Generating classification weights with GNN denoising autoencoders for few-shot learning[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2019：21-30.
[4] TAO X，HONG X，SHI W，et al.Analogy-detail networks for object recognition[J].IEEE Transactions on Neural Networks and Learning Systems，2021，32：4404-4418.
[5] RAJASEGARAN J，KHAN S，HAYAT M，et al.iTAML：an incremental task-agnostic meta-learning approach[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2020：13585-13594.
[6] YU L，TWARDOWSKI B，LIU X，et al.Semantic drift compensation for class-incremental learning[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2020：6980-6989.
[7] CHERAGHIAN A，RAHMAN S，FANG P.Semantic-aware knowledge distillation for few-shot class-incremental learning[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2021：2534-2543.
[8] 韩纪东，李玉鑑.神经网络模型中灾难性遗忘研究的综述[J].北京工业大学学报，2021，47（5）：551-564.
HAN J D，LI Y J.Survey of catastrophic forgetting research in neural network models[J].Journal of Beijing University of Technology.2021，47（5）：551-564.
[9] REBUFFI S A，KOLESNIKOV A，SPERL G，et al.iCaRL：incremental classifier and representation learning[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2017：5533-5542.
[10] HOU S，PAN X，LOY C C，et al.Learning a unified classifier incrementally via rebalancing[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2019：831-839.
[11] KRIZHEVSKY A，HINTON G.Learning multiple layers of features from tiny images[J].Handbook of Systemic Autoimmune Diseases，2009，1（4）.
[12] VINYALS O，BLUNDELL C，LILLICRAP T，et al.Matching networks for one shot learning[C]//Advances in Neural Information Processing Systems，2016：3630-3638.
[13] LECUN Y，BENGIO Y，HINTON G.Deep learning[J].Nature，2015，521（7553）：436-444.
[14] HE K，ZHANG X，REN S，et al.Deep residual learning for image recognition[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2016：770-778.
[15] RUSSAKOVSKY O，DENG J，SU H，et al.Imagenet large scale visual recognition challenge[J].International Journal of Computer Vision，2015，115（3）：211-252.
[16] CASTRO F M，MAR′IN-JIM′ENEZ M J，GUIL N，et al.End-to-end incremental learning[C]//European Conference on Computer Vision，2018.
[17] 莫建文，陈瑶嘉.基于VAE的伪样本重排练实现的类增量学习[J].计算机工程与设计，2021，42（8）：2284-2290.
MO J W，CHEN Y J.Class incremental learning based on pseudo-rehearsal with variational autoencoder[J].Computer Engineering and Design，2021，42（8）：2284-2290.
[18] LI Z，HOIEM D.Learning without forgetting[J].IEEE Transactions on Pattern Analysis and Machine Intelligence，2018，40（12）：2935-2947.
[19] KIRKPATRICK J，PASCANU R，RABINOWITZ N，et al.Overcoming catastrophic forgetting in neural networks[J].Proceedings of the National Academy of Sciences of the United States of America，2017，114：3521-3526.
[20] REN M，LIAO R，FETAYA E，et al.Incremental few-shot learning with attention attractor networks[C]//Advances in Neural Information Processing Systems，2019.
[21] GIDARIS S，KOMODAKIS N.Dynamic few-shot visual learning without forgetting[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2018：4367-4375.