Is PageRank All You Need for Scalable Graph Neural Networks?

Is PageRank All You Need for Scalable Graph Neural Networks?. Bojchevski, A., Klicpera, J., Perozzi, B., Blais, M., Kapoor, A., Lukasik, M., & Günnemann, S. 2019.
abstract bibtex

Graph neural networks (GNNs) have emerged as a powerful approach for solving many network mining tasks. However, efficiently utilizing them on web-scale data remains a challenge despite related advances in research. Most recently proposed scalable GNNs rely on an expensive recursive message-passing procedure to propagate information through the graph. We circumvent this limitation by leveraging connections between GNNs and personalized PageRank and we develop a model that incorporates multi-hop neighborhood information in a single (non-recursive) step. Our work-in-progress approach PPRGo is significantly faster than multi-hop models while maintaining state-of-the-art prediction performance. We demonstrate the strengths and scalability of our approach on graphs orders of magnitude larger than typically considered in the literature.

@article{bojchevski_is_2019,
	title = {Is {PageRank} {All} {You} {Need}  for {Scalable} {Graph} {Neural} {Networks}?},
	abstract = {Graph neural networks (GNNs) have emerged as a powerful approach for solving many network mining tasks. However, efficiently utilizing them on web-scale data remains a challenge despite related advances in research. Most recently proposed scalable GNNs rely on an expensive recursive message-passing procedure to propagate information through the graph. We circumvent this limitation by leveraging connections between GNNs and personalized PageRank and we develop a model that incorporates multi-hop neighborhood information in a single (non-recursive) step. Our work-in-progress approach PPRGo is significantly faster than multi-hop models while maintaining state-of-the-art prediction performance. We demonstrate the strengths and scalability of our approach on graphs orders of magnitude larger than typically considered in the literature.},
	language = {en},
	author = {Bojchevski, Aleksandar and Klicpera, Johannes and Perozzi, Bryan and Blais, Martin and Kapoor, Amol and Lukasik, Michal and Günnemann, Stephan},
	year = {2019},
	keywords = {/unread, ⛔ No DOI found},
	pages = {7},
}

Downloads: 0

{"_id":"C79xLjDgc2JoDoGLn","bibbaseid":"bojchevski-klicpera-perozzi-blais-kapoor-lukasik-gnnemann-ispagerankallyouneedforscalablegraphneuralnetworks-2019","author_short":["Bojchevski, A.","Klicpera, J.","Perozzi, B.","Blais, M.","Kapoor, A.","Lukasik, M.","Günnemann, S."],"bibdata":{"bibtype":"article","type":"article","title":"Is PageRank All You Need for Scalable Graph Neural Networks?","abstract":"Graph neural networks (GNNs) have emerged as a powerful approach for solving many network mining tasks. However, efficiently utilizing them on web-scale data remains a challenge despite related advances in research. Most recently proposed scalable GNNs rely on an expensive recursive message-passing procedure to propagate information through the graph. We circumvent this limitation by leveraging connections between GNNs and personalized PageRank and we develop a model that incorporates multi-hop neighborhood information in a single (non-recursive) step. Our work-in-progress approach PPRGo is significantly faster than multi-hop models while maintaining state-of-the-art prediction performance. We demonstrate the strengths and scalability of our approach on graphs orders of magnitude larger than typically considered in the literature.","language":"en","author":[{"propositions":[],"lastnames":["Bojchevski"],"firstnames":["Aleksandar"],"suffixes":[]},{"propositions":[],"lastnames":["Klicpera"],"firstnames":["Johannes"],"suffixes":[]},{"propositions":[],"lastnames":["Perozzi"],"firstnames":["Bryan"],"suffixes":[]},{"propositions":[],"lastnames":["Blais"],"firstnames":["Martin"],"suffixes":[]},{"propositions":[],"lastnames":["Kapoor"],"firstnames":["Amol"],"suffixes":[]},{"propositions":[],"lastnames":["Lukasik"],"firstnames":["Michal"],"suffixes":[]},{"propositions":[],"lastnames":["Günnemann"],"firstnames":["Stephan"],"suffixes":[]}],"year":"2019","keywords":"/unread, ⛔ No DOI found","pages":"7","bibtex":"@article{bojchevski_is_2019,\n\ttitle = {Is {PageRank} {All} {You} {Need} for {Scalable} {Graph} {Neural} {Networks}?},\n\tabstract = {Graph neural networks (GNNs) have emerged as a powerful approach for solving many network mining tasks. However, efficiently utilizing them on web-scale data remains a challenge despite related advances in research. Most recently proposed scalable GNNs rely on an expensive recursive message-passing procedure to propagate information through the graph. We circumvent this limitation by leveraging connections between GNNs and personalized PageRank and we develop a model that incorporates multi-hop neighborhood information in a single (non-recursive) step. Our work-in-progress approach PPRGo is significantly faster than multi-hop models while maintaining state-of-the-art prediction performance. We demonstrate the strengths and scalability of our approach on graphs orders of magnitude larger than typically considered in the literature.},\n\tlanguage = {en},\n\tauthor = {Bojchevski, Aleksandar and Klicpera, Johannes and Perozzi, Bryan and Blais, Martin and Kapoor, Amol and Lukasik, Michal and Günnemann, Stephan},\n\tyear = {2019},\n\tkeywords = {/unread, ⛔ No DOI found},\n\tpages = {7},\n}\n\n","author_short":["Bojchevski, A.","Klicpera, J.","Perozzi, B.","Blais, M.","Kapoor, A.","Lukasik, M.","Günnemann, S."],"key":"bojchevski_is_2019","id":"bojchevski_is_2019","bibbaseid":"bojchevski-klicpera-perozzi-blais-kapoor-lukasik-gnnemann-ispagerankallyouneedforscalablegraphneuralnetworks-2019","role":"author","urls":{},"keyword":["/unread","⛔ No DOI found"],"metadata":{"authorlinks":{}},"html":""},"bibtype":"article","biburl":"https://bibbase.org/zotero/victorjhu","dataSources":["CmHEoydhafhbkXXt5"],"keywords":["/unread","⛔ no doi found"],"search_terms":["pagerank","need","scalable","graph","neural","networks","bojchevski","klicpera","perozzi","blais","kapoor","lukasik","günnemann"],"title":"Is PageRank All You Need for Scalable Graph Neural Networks?","year":2019}