“…T HERE are ubiquitous graph structures in various realworld complex systems, which call for trustworthy and effective graph characterization paradigms for more accurate and ultimately more useful representations. During the past decades, dozens of graph representation methods have been proposed for many networks, including communication networks [1], [2], social networks [3], [4], and biological networks [5], [6], etc. Generally, representation learning for networks is widely considered as a promising yet more challenging task, which This work was supported in part by the National Natural Science Foundation of China under Grant 61673178 and 61922063; in part by the Natural Science Foundation of Shanghai under Grant 20ZR1413800; in part by the European Union's Horizon 2020 research and innovation programme under the Marie Sklodowska-Curie grant agreement No 824019 and 101022280.…”