アブストラクト
| Date | 14:55-15:20 Jan 30, 2024 |
| Speaker | Zhang Yue School of Computer Science, Guangdong Polytechnic Normal University, China |
| Title | Collaborative embedding learning via tensor integration for multi-view clustering |
| Abstract |
Multi-view clustering takes advantage of diverse perspectives to enable
comprehensive data analysis. Recently, graph learning via
low-dimensional embedding has emerged in multi-view clustering to learn
a consensus affinity graph. Despite its efficiency, projecting data into
the low-dimensional space has often resulted in the compression of data
information. To overcome this challenge, this paper proposes a
Collaborative Embedding Learning via Tensor (CELT) method. CELT
simultaneously learns intra-view affinity graphs for each view from both
the original and low-dimensional spaces. Additionally, all intra-view
affinity graphs are stacked into an overcomplete affinity tensor,
enabling an enhanced consensus affinity across inter-view consistency.
Extensive experiments demonstrate that the proposed collaborative
learning framework effectively improves graph learning and outperforms
competitive multi-view clustering methods.
|