多角度语义标签引导的自监督多视图聚类  

Multi-view clustering with self-supervised learningguided by multi-angle semantic labels

作　　者：柳源 安俊秀[1,2] 杨林旺 Liu Yuan;An Junxiu;Yang Linwang(School of Software Engineering,Chengdu University of Information Technology,Chengdu 610000,China;Key Laboratory of Manufactu-ring Industry Chain Collaboration&Information Technology Support,Sichuan Province,Chengdu 610000,China)

机构地区：[1]成都信息工程大学软件工程学院,成都610000 [2]制造业产业链协同与信息化支撑技术四川省重点实验室,成都610000

出　　处：《计算机应用研究》2024年第11期3357-3363,共7页Application Research of Computers

基　　金：国家社会科学基金资助项目(22BXW048);四川省重点实验室开放基金资助项目(2024-ScL-MC&I-001);成都市科技重点研发支撑计划资助项目(2022-YF05-00454-SN)。

摘　　要：多视图聚类旨在从多个角度挖掘对象的特征信息,以获得精准的聚类结果。然而,现有研究往往无法妥善处理视图融合时产生的信息冲突,并且对多视图之间的互补信息利用不够充分。为解决这些问题,提出了一种由多角度语义标签引导的自监督多视图聚类模型。该模型首先将各视图的潜在表示映射到独立的低维特征空间,在一个空间中专注于优化视图间的一致性,以维护特征空间的局部结构和样本间的相对关系;同时,在另一空间中直接从视图层面提取聚类信息,以捕获更丰富多样的语义特征;最后,利用多个角度语义特征生成的伪标签,引导对象层面的聚类分配,实现两种表示的协同优化。大量实验结果表明,该方法能够全面挖掘多视图数据中的公共信息与互补信息,并展现出良好的聚类性能。此外,相较于其他方法,该方法在视图数量较多的场景更具优势。Multi-view clustering aims to explore the feature information of objects from multiple perspectives to obtain accurate clustering results.However,existing research often fails to handle the information conflicts that arise during view fusion and does not fully utilize the complementary information between multiple views.To address these issues,this paper proposed a self-supervised multi-view clustering model guided by multi-angle semantic labels.The model first mapped the latent representations of each view to independent low-dimensional feature spaces,focusing on optimizing the consistency between views in one space to maintain the local structure of the feature space and the relative relationships between samples.At the same time,in another space,clustering information was directly extracted from the view level to capture richer and more diverse semantic features.Finally,pseudo-labels generated from multi-angle semantic features guided the clustering assignment at the object level,achieving collaborative optimization of the two representations.Extensive experimental results demonstrate that this approach can comprehensively explore both common and complementary information in multi-view data and exhibit good clustering performance.Moreover,compared to other methods,this approach has advantages in scenarios with a larger number of views.

关 键 词：多视图聚类 无监督学习 对比学习 深度聚类 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]