{"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/dagcn-dual-attention-graph-convolutional","title":"DAGCN: Dual Attention Graph Convolutional Networks","arxiv_id":"1904.02278","date":"2019-04-04","proceeding":null,"authors":["Fengwen Chen","Shirui Pan","Jing Jiang","Huan Huo","Guodong Long"],"abstract":"Graph convolutional networks (GCNs) have recently become one of the most\npowerful tools for graph analytics tasks in numerous applications, ranging from\nsocial networks and natural language processing to bioinformatics and\nchemoinformatics, thanks to their ability to capture the complex relationships\nbetween concepts. At present, the vast majority of GCNs use a neighborhood\naggregation framework to learn a continuous and compact vector, then performing\na pooling operation to generalize graph embedding for the classification task.\nThese approaches have two disadvantages in the graph classification task:\n(1)when only the largest sub-graph structure ($k$-hop neighbor) is used for\nneighborhood aggregation, a large amount of early-stage information is lost\nduring the graph convolution step; (2) simple average/sum pooling or max\npooling utilized, which loses the characteristics of each node and the topology\nbetween nodes. In this paper, we propose a novel framework called, dual\nattention graph convolutional networks (DAGCN) to address these problems. DAGCN\nautomatically learns the importance of neighbors at different hops using a\nnovel attention graph convolution layer, and then employs a second attention\ncomponent, a self-attention pooling layer, to generalize the graph\nrepresentation from the various aspects of a matrix graph embedding. The dual\nattention network is trained in an end-to-end manner for the graph\nclassification task. We compare our model with state-of-the-art graph kernels\nand other deep learning methods. The experimental results show that our\nframework not only outperforms other baselines but also achieves a better rate\nof convergence.","url_abs":"http://arxiv.org/abs/1904.02278v1","url_pdf":"http://arxiv.org/pdf/1904.02278v1.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":"dagcn-dual-attention-graph-convolutional","repo_url":"https://github.com/dawenzi123/DAGCN","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"classification","task_name":"General Classification"},{"task_slug":"graph-classification","task_name":"Graph Classification"},{"task_slug":"graph-embedding","task_name":"Graph Embedding"}],"methods":[{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"graph-convolutional-networks","method_name":"Graph Convolutional Networks"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/graph-classification-on-enzymes","task":"Graph Classification","dataset":"ENZYMES","model":"DAGCN","rank_in_archive_order":35,"of":54,"metrics":{"Accuracy":"58.17%"},"uses_additional_data":false},{"leaderboard":"/sota/graph-classification-on-mutag","task":"Graph Classification","dataset":"MUTAG","model":"DAGCN","rank_in_archive_order":49,"of":74,"metrics":{"Accuracy":"87.22%"},"uses_additional_data":false},{"leaderboard":"/sota/graph-classification-on-nci1","task":"Graph Classification","dataset":"NCI1","model":"DAGCN","rank_in_archive_order":33,"of":69,"metrics":{"Accuracy":"81.68%"},"uses_additional_data":false},{"leaderboard":"/sota/graph-classification-on-proteins","task":"Graph Classification","dataset":"PROTEINS","model":"DAGCN","rank_in_archive_order":50,"of":103,"metrics":{"Accuracy":"76.33%"},"uses_additional_data":false},{"leaderboard":"/sota/graph-classification-on-ptc","task":"Graph Classification","dataset":"PTC","model":"DAGCN","rank_in_archive_order":28,"of":37,"metrics":{"Accuracy":"62.88%"},"uses_additional_data":false}],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}