电子学报2025，Vol.53Issue(11)：3970-3982,13.DOI:10.12263/DZXB.20250413
面向在线连续学习的特征融合引导的梯度重加权算法
Gradient Re-Weighting Guided by Feature Fusion for Online Continual Learning
摘要
Abstract
Online continual learning(OCL)aims at learning a non-stationary data stream in a way of reading each da-ta sample only once,and hence suffers from insufficient learning.To address this problem,we propose a feature fusion method in this work.Our method leverages augmented samples of an image for producing anchor features,and incorporates them to obtain a fused feature via a weighted summation operation.The weights are determined by the similarity between anchor features and a pre-designated pivotal feature of the image.Optimizing the cross-entropy loss of this fused feature can accelerate the learning process,resulting in better performance on the current task.Additionally,we propose a consistency loss that restricts the mean-square error between the fused feature and the pivotal feature,which can further improve the per-formance on the current task.Finally,we provide a theoretical analysis about the gradients of cross-entropy loss to model parameters.This analysis reveals the relationship between the feature fusion and the gradient re-weighting.Extensive experi-ments are conducted on three benchmarks under OCL settings,including CIFAR-10,CIFAR-100 and Tiny-ImageNet.Our method surpasses baselines at most 7.00%,8.04%,6.33%for average end accuracy on CIFAR-10,CIFAR-100 and Tiny-Im-ageNet,respectively.Experimental results demonstrate the proposed method is effective,and achieves substantial improve-ment over previous methods for online continual learning.关键词
图像识别/连续学习/在线学习/类别增量学习/特征融合/梯度重加权Key words
image recognition/continual learning/online learning/class incremental learning/feature fusion/gradi-ent re-weighting分类
信息技术与安全科学引用本文复制引用
邱奔流,王岚晓,邱荷茜,高翔宇,问海涛,李宏亮..面向在线连续学习的特征融合引导的梯度重加权算法[J].电子学报,2025,53(11):3970-3982,13.基金项目
新一代人工智能国家科技重大专项(No.2021ZD0112001) National Science and Technology Major Project(No.2021ZD0112001) （No.2021ZD0112001）
