{"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/deeply-learning-molecular-structure-property","title":"Deeply learning molecular structure-property relationships using attention- and gate-augmented graph convolutional network","arxiv_id":"1805.10988","date":"2018-05-28","proceeding":null,"authors":["Seongok Ryu","Jaechang Lim","Seung Hwan Hong","Woo Youn Kim"],"abstract":"Molecular structure-property relationships are key to molecular engineering\nfor materials and drug discovery. The rise of deep learning offers a new viable\nsolution to elucidate the structure-property relationships directly from\nchemical data. Here we show that the performance of graph convolutional\nnetworks (GCNs) for the prediction of molecular properties can be improved by\nincorporating attention and gate mechanisms. The attention mechanism enables a\nGCN to identify atoms in different environments. The gated skip-connection\nfurther improves the GCN by updating feature maps at an appropriate rate. We\ndemonstrate that the resulting attention- and gate-augmented GCN could extract\nbetter structural features related to a target molecular property such as\nsolubility, polarity, synthetic accessibility and photovoltaic efficiency\ncompared to the vanilla GCN. More interestingly, it identified two distinct\nparts of molecules as essential structural features for high photovoltaic\nefficiency, and each of them coincided with the areas of donor and acceptor\norbitals for charge-transfer excitations, respectively. As a result, the new\nmodel could accurately predict molecular properties and place molecules with\nsimilar properties close to each other in a well-trained latent space, which is\ncritical for successful molecular engineering.","url_abs":"http://arxiv.org/abs/1805.10988v3","url_pdf":"http://arxiv.org/pdf/1805.10988v3.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":"deeply-learning-molecular-structure-property","repo_url":"https://github.com/qyuan7/molecular_gcn","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"drug-discovery","task_name":"Drug Discovery"}],"methods":[{"method_slug":"gcn","method_name":"GCN"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1805.10988","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}