文献类型：期刊文献

英文题名：Multi-view prototype balance and temporary proxy constraint for exemplar-free class-incremental learning

作者：Tian, Heng[1,2];Zhang, Qian[1,2];Wang, Zhe[1,2];Zhang, Yu[1,2];Xu, Xinlei[1,2];Fu, Zhiling[1,2]

机构：[1]Minist Educ, Key Lab Smart Mfg Energy Chem Proc, Shanghai 200237, Peoples R China;[2]East China Univ Sci & Technol, Dept Comp Sci & Engn, Shanghai 200237, Peoples R China

年份：2025

卷号：55

期号：5

外文期刊名：APPLIED INTELLIGENCE

收录：;EI(收录号：20250717889256);WOS:【SCI-EXPANDED(收录号:WOS:001399516800003)】；

基金：This work is supported by Natural Science Foundation of China under Grant No. 62476087, the National Key Research and Development Program of China (2022YFB3203500), Shanghai Science and Technology Program "Federated based cross-domain and cross-task incremental learning" under Grant No. 21511100800, Natural Science Foundation of China under Grant No. 62002193, Chinese Defense Program of Science and Technology under Grant No.2021-JCJQ-JJ-0041, China Aerospace Science and Technology Corporation Industry-University-Research Cooperation Foundation of the Eighth Research Institute under Grant No.SAST2021-007.

语种：英文

外文关键词：Class-incremental learning; Exemplar-free; Prototype balance; Proxy constraint; Multi-view

摘要：Exemplar-free class-incremental learning recognizes both old and new classes without saving old class exemplars because of storage limitations and privacy constraints. To address the forgetting of knowledge caused by the absence of old training data, we present a novel method that consists of two modules, multi-view prototype balance and temporary proxy constraints, which are based on feature retention and representation optimization. Specifically, multi-view prototype balance first extends the prototypes to maintain the general state of the class and then balances these prototypes combining knowledge distillation and prototype compensation to ensure the stability and plasticity of the model. To alleviate the feature overlap, the proposed temporary proxy constraint sets the temporary proxies to lightly compress the feature distribution during each mini-batch of training. Extensive experiments on five datasets with different settings demonstrate the superiority of our method against the state-of-the-art exemplar-free class-incremental learning methods.