基于多原型重放和对齐的类增量无源域适应  

Class-incremental Source-free Domain Adaptation Based on Multi-prototype Replay and Alignment

作　　者：田青 康陆禄 周亮宇 TIAN Qing;KANG Lulu;ZHOU Liangyu(School of Software,Nanjing University of Information Science and Technology,Nanjing 210044,China;Wuxi Institute of Technology,Nanjing University of Information Science and Technology,Wuxi,Jiangsu 214000,China;State Key Laboratory for Novel Software Technology,Nanjing University,Nanjing 210023,China)

机构地区：[1]南京信息工程大学软件学院,南京210044 [2]南京信息工程大学无锡研究院,江苏无锡214000 [3]南京大学计算机软件新技术国家重点实验室,南京210023

出　　处：《计算机科学》2025年第3期206-213,共8页Computer Science

基　　金：国家自然科学基金(62176128);江苏省自然科学基金(BK20231143);南京大学计算机软件新技术国家重点实验室开放课题(KFKT2022B06);中央高校基本科研基金(NJ2022028);江苏省“青蓝工程”人才计划项目。

摘　　要：传统无源域适应通常假设目标域数据全部可用,然而在实际应用中目标域数据常以流的形式出现,即未标记的目标域中的类会依次增加,这无疑带来了新的挑战。首先,在每个时间步骤中,目标域的标签空间都是源域的一个子集,盲目对齐反而会导致模型性能下降;其次,在学习新类的过程中会破坏先前学习到的知识,导致之前知识的灾难性遗忘。为了解决这些问题,提出了一种基于多原型重放和对齐(MPRA)的方法。该方法通过累积预测概率检测目标域中的共享类来应对标签空间不一致问题,并采用多原型重放来处理灾难性遗忘,提高模型的记忆能力。同时,基于多原型和源模型权重进行跨域的对比学习,从而对齐特征分布,提高模型性能。大量的实验表明,所提方法在3个基准数据集上都取得了优越的表现。Traditional source-free domain adaptation usually assumes that all the target domain data is available,but in practice,the target domain data often appears in the form of streams,that is,the classes in the unlabeled target domain will increase sequentially,which undoubtedly brings new challenges.First,in each time step,the label space of the target domain is a subset of the source domain,and blind alignment will cause the performance of the model to deteriorate.Secondly,in the process of learning new classes,it will destroy the previously learned knowledge,resulting in the catastrophic forgetting of the previous knowledge.In order to solve these problems,this paper proposes a method based on multi-prototype replay and alignment(MPRA).In this method,the shared classes in the target domain are detected by cumulative prediction probabilities,the label space inconsistency problem is solved,and the multi-prototype replay is used to deal with catastrophic forgetting and improve the memory ability of the model.Additionally,the method incorporates cross-domain contrastive learning based on multi-prototype and source model weights to align feature distributions and improve model robustness.A large number of experiments show that the proposed method has achieved superior performance on 3 benchmark datasets.

关 键 词：无源域适应 类增量学习 多原型 对比学习 迁移学习 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]