{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/geometric-matrix-completion-with-recurrent","title":"Geometric Matrix Completion with Recurrent Multi-Graph Neural Networks","arxiv_id":"1704.06803","date":"2017-04-22","proceeding":"NeurIPS 2017 12","authors":["Federico Monti","Michael M. Bronstein","Xavier Bresson"],"abstract":"Matrix completion models are among the most common formulations of\nrecommender systems. Recent works have showed a boost of performance of these\ntechniques when introducing the pairwise relationships between users/items in\nthe form of graphs, and imposing smoothness priors on these graphs. However,\nsuch techniques do not fully exploit the local stationarity structures of\nuser/item graphs, and the number of parameters to learn is linear w.r.t. the\nnumber of users and items. We propose a novel approach to overcome these\nlimitations by using geometric deep learning on graphs. Our matrix completion\narchitecture combines graph convolutional neural networks and recurrent neural\nnetworks to learn meaningful statistical graph-structured patterns and the\nnon-linear diffusion process that generates the known ratings. This neural\nnetwork system requires a constant number of parameters independent of the\nmatrix size. We apply our method on both synthetic and real datasets, showing\nthat it outperforms state-of-the-art techniques.","url_abs":"http://arxiv.org/abs/1704.06803v1","url_pdf":"http://arxiv.org/pdf/1704.06803v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"geometric-matrix-completion-with-recurrent","repo_url":"https://github.com/fmonti/mgcnn","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}},{"paper_slug":"geometric-matrix-completion-with-recurrent","repo_url":"https://github.com/zuirod/mg-gat","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"tf","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":"collaborative-filtering","task_name":"Collaborative Filtering"},{"task_slug":"matrix-completion","task_name":"Matrix Completion"},{"task_slug":"recommendation-systems","task_name":"Recommendation Systems"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/recommendation-systems-on-douban-monti","task":"Recommendation Systems","dataset":"Douban Monti","model":"sRGCNN","rank_in_archive_order":7,"of":8,"metrics":{"RMSE":"0.8012"},"uses_additional_data":true},{"leaderboard":"/sota/recommendation-systems-on-flixster-monti","task":"Recommendation Systems","dataset":"Flixster Monti","model":"sRGCNN","rank_in_archive_order":6,"of":7,"metrics":{"RMSE":"0.9258"},"uses_additional_data":true},{"leaderboard":"/sota/collaborative-filtering-on-movielens-100k","task":"Recommendation Systems","dataset":"MovieLens 100K","model":"sRGCNN","rank_in_archive_order":14,"of":18,"metrics":{"RMSE (u1 Splits)":"0.929"},"uses_additional_data":true},{"leaderboard":"/sota/recommendation-systems-on-yahoomusic-monti","task":"Recommendation Systems","dataset":"YahooMusic Monti","model":"sRGCNN","rank_in_archive_order":5,"of":6,"metrics":{"RMSE":"22.4149"},"uses_additional_data":true}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1704.06803","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}