On a linear fused Gromov-Wasserstein distance for graph structured data

Open Webpage

Dai Hai Nguyen, Koji Tsuda

Published: 01 Jan 2023, Last Modified: 27 Sept 2024Pattern Recognit. 2023EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•An optimal transport-based distance, named linearFGW, is proposed for learning with graph structured data.•Theoretical properties of the proposed linearFGW are provided.•Experimental results on graph classification and clustering tasks are demonstrated, showing the effectiveness of the linearFGW.