{"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/chemixnet-mixed-dnn-architectures-for","title":"CheMixNet: Mixed DNN Architectures for Predicting Chemical Properties using Multiple Molecular Representations","arxiv_id":"1811.08283","date":"2018-11-14","proceeding":null,"authors":["Arindam Paul","Dipendra Jha","Reda Al-Bahrani","Wei-keng Liao","Alok Choudhary","Ankit Agrawal"],"abstract":"SMILES is a linear representation of chemical structures which encodes the\nconnection table, and the stereochemistry of a molecule as a line of text with\na grammar structure denoting atoms, bonds, rings and chains, and this\ninformation can be used to predict chemical properties. Molecular fingerprints\nare representations of chemical structures, successfully used in similarity\nsearch, clustering, classification, drug discovery, and virtual screening and\nare a standard and computationally efficient abstract representation where\nstructural features are represented as a bit string. Both SMILES and molecular\nfingerprints are different representations for describing the structure of a\nmolecule. There exist several predictive models for learning chemical\nproperties based on either SMILES or molecular fingerprints. Here, our goal is\nto build predictive models that can leverage both these molecular\nrepresentations. In this work, we present CheMixNet -- a set of neural networks\nfor predicting chemical properties from a mixture of features learned from the\ntwo molecular representations -- SMILES as sequences and molecular fingerprints\nas vector inputs. We demonstrate the efficacy of CheMixNet architectures by\nevaluating on six different datasets. The proposed CheMixNet models not only\noutperforms the candidate neural architectures such as contemporary fully\nconnected networks that uses molecular fingerprints and 1-D CNN and RNN models\ntrained SMILES sequences, but also other state-of-the-art architectures such as\nChemception and Molecular Graph Convolutions.","url_abs":"http://arxiv.org/abs/1811.08283v2","url_pdf":"http://arxiv.org/pdf/1811.08283v2.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":"chemixnet-mixed-dnn-architectures-for","repo_url":"https://github.com/paularindam/CheMixNet","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":null},{"paper_slug":"chemixnet-mixed-dnn-architectures-for","repo_url":"https://github.com/CAVED123/CheMixNet","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null},{"paper_slug":"chemixnet-mixed-dnn-architectures-for","repo_url":"https://github.com/nu-cucis/chemixnet","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":"clustering","task_name":"Clustering"},{"task_slug":"drug-discovery","task_name":"Drug Discovery"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}