arrow
Return

Mixed-Modality Clustering via Generative Graph Structure Matching

delete2024-12-01
delete0
delete
OA
AI
X
Xiaxia He
王博岳 (Boyue Wang) *
J
Junbin Gao
Q
Qianqian Wang
Y
Yongli Hu
B
Baocai Yin
DOI:10.1109/TKDE.2024.3434556delete
deleteOriginal
deleteOriginal request for help
deleteShare
deleteSave
Abstract

Abstract

En 中文
The goal of mixed-modality clustering, which differs from typical multi-modality/view clustering, is to divide samples derived from various modalities into several clusters. This task has to solve two critical semantic gap problems: i) how to generate the missing modalities without the pairwise-modality data; and ii) how to align the representations of heterogeneous modalities. To tackle the above problems, this paper proposes a novel mixed-modality clustering model, which integrates the missing-modality generation and the heterogeneous modality alignment into a unified framework. During the missing-modality generation process, a bidirectional mapping is established between different modalities, enabling generation of preliminary representations for the missing-modality using information from another modality. Then the intra-modality bipartite graphs are constructed to help generate better missing-modality representations by weighted aggregating existing intra-modality neighbors. In this way, a pairwise-modality representation for each sample can be obtained. In the process of heterogeneous modality alignment, each modality is modelled as a graph to capture the global structure among intra-modality samples and is aligned against the heterogeneous modality representations through the adaptive heterogeneous graph matching module. Experimental results on three public datasets show the effectiveness of the proposed model compared to multiple state-of-the-art multi-modality/view clustering methods.
Keywords:
Correlation
Semantics
Task analysis
Data models
Web sites
Feature extraction
Bipartite graph
Mixed-modality clustering
multi-view clustering
adaptive graph structure learning
heterogeneous graph matching

Journal

IEEE Transactions on Knowledge and Data Engineering cover
IEEE Transactions on Knowledge and Data Engineering
IF:
10.4
Papers:
6.7K
Citations:
3.2W

Organization

U
University of Sydney
Scholars:
6.5W
Papers: 6.2W
Citations: 90
B
Beijing University of Technology
Scholars:
2.8W
Papers: 2.1W
Citations: 2.7W
X
Xidian University
Scholars:
2.4W
Papers: 1.9W
Citations: 9.7K
researcher View more organizations