Graph Distance

Graph distance measures quantify the dissimilarity between graph structures, a crucial task in various fields requiring analysis of relational data. Current research focuses on developing efficient and accurate graph distance metrics, including those based on spectral properties, edit distances, and Gromov-Wasserstein distances, as well as exploring their use in evaluating graph neural network performance and generating diverse graph datasets. These advancements are improving the analysis of complex networks in diverse applications, such as molecular property prediction, semantic graph parsing, and semi-supervised learning.