{"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/learning-graph-pooling-and-hybrid","title":"Learning Graph Pooling and Hybrid Convolutional Operations for Text Representations","arxiv_id":"1901.06965","date":"2019-01-21","proceeding":null,"authors":["Hongyang Gao","Yongjun Chen","Shuiwang Ji"],"abstract":"With the development of graph convolutional networks (GCN), deep learning\nmethods have started to be used on graph data. In additional to convolutional\nlayers, pooling layers are another important components of deep learning.\nHowever, no effective pooling methods have been developed for graphs currently.\nIn this work, we propose the graph pooling (gPool) layer, which employs a\ntrainable projection vector to measure the importance of nodes in graphs. By\nselecting the k-most important nodes to form the new graph, gPool achieves the\nsame objective as regular max pooling layers operating on images. Another\nlimitation of GCN when used on graph-based text representation tasks is that,\nGCNs do not consider the order information of nodes in graph. To address this\nlimitation, we propose the hybrid convolutional (hConv) layer that combines GCN\nand regular convolutional operations. The hConv layer is capable of increasing\nreceptive fields quickly and computing features automatically. Based on the\nproposed gPool and hConv layers, we develop new deep networks for text\ncategorization tasks. Our results show that the networks based on gPool and\nhConv layers achieves new state-of-the-art performance as compared to baseline\nmethods.","url_abs":"http://arxiv.org/abs/1901.06965v2","url_pdf":"http://arxiv.org/pdf/1901.06965v2.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":"learning-graph-pooling-and-hybrid","repo_url":"https://github.com/HongyangGao/hConv-gPool-Net","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":"text-categorization","task_name":"Text Categorization"}],"methods":[{"method_slug":"gcn","method_name":"GCN"},{"method_slug":"graph-convolutional-networks","method_name":"Graph Convolutional Networks"},{"method_slug":"max-pooling","method_name":"Max Pooling"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1901.06965","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}