{"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/schnet-a-continuous-filter-convolutional","title":"SchNet: A continuous-filter convolutional neural network for modeling quantum interactions","arxiv_id":"1706.08566","date":"2017-06-26","proceeding":"NeurIPS 2017 12","authors":["Kristof T. Schütt","Pieter-Jan Kindermans","Huziel E. Sauceda","Stefan Chmiela","Alexandre Tkatchenko","Klaus-Robert Müller"],"abstract":"Deep learning has the potential to revolutionize quantum chemistry as it is\nideally suited to learn representations for structured data and speed up the\nexploration of chemical space. While convolutional neural networks have proven\nto be the first choice for images, audio and video data, the atoms in molecules\nare not restricted to a grid. Instead, their precise locations contain\nessential physical information, that would get lost if discretized. Thus, we\npropose to use continuous-filter convolutional layers to be able to model local\ncorrelations without requiring the data to lie on a grid. We apply those layers\nin SchNet: a novel deep learning architecture modeling quantum interactions in\nmolecules. We obtain a joint model for the total energy and interatomic forces\nthat follows fundamental quantum-chemical principles. This includes\nrotationally invariant energy predictions and a smooth, differentiable\npotential energy surface. Our architecture achieves state-of-the-art\nperformance for benchmarks of equilibrium molecules and molecular dynamics\ntrajectories. Finally, we introduce a more challenging benchmark with chemical\nand structural variations that suggests the path for further work.","url_abs":"http://arxiv.org/abs/1706.08566v5","url_pdf":"http://arxiv.org/pdf/1706.08566v5.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":"schnet-a-continuous-filter-convolutional","repo_url":"https://github.com/atomistic-machine-learning/SchNet","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"tf","reach":{"status":"unanswered"}},{"paper_slug":"schnet-a-continuous-filter-convolutional","repo_url":"https://github.com/atomistic-machine-learning/schnetpack","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"ok","spdx":"NOASSERTION"}},{"paper_slug":"schnet-a-continuous-filter-convolutional","repo_url":"https://github.com/jmg764/Molecular-Scalar-Coupling-Constant-Prediction-using-SchNet","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"unanswered"}},{"paper_slug":"schnet-a-continuous-filter-convolutional","repo_url":"https://github.com/jmg764/Predicting-Molecular-Scalar-Couplings-using-SchNet","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"unanswered"}},{"paper_slug":"schnet-a-continuous-filter-convolutional","repo_url":"https://github.com/xiuyu0000/new_papers_codes/tree/main/Cybertron","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"mindspore","reach":null}],"tasks":[{"task_slug":"formation-energy","task_name":"Formation Energy"},{"task_slug":"time-series-1","task_name":"Time Series"}],"methods":[{"method_slug":"speed","method_name":"SPEED"},{"method_slug":"schnet","method_name":"SchNet"},{"method_slug":"ssp","method_name":"Shifted Softplus"}],"datasets_introduced":[],"methods_introduced":[{"slug":"schnet","name":"SchNet","full_name":"Schrödinger Network"},{"slug":"ssp","name":"Shifted Softplus","full_name":"Shifted Softplus"}],"results":[{"leaderboard":"/sota/formation-energy-on-jarvis-dft-formation","task":"Formation Energy","dataset":"JARVIS-DFT","model":"SchNet","rank_in_archive_order":5,"of":6,"metrics":{"MAE":"0.045"},"uses_additional_data":false},{"leaderboard":"/sota/formation-energy-on-oqm9hk","task":"Formation Energy","dataset":"OQM9HK","model":"SchNet","rank_in_archive_order":4,"of":4,"metrics":{"MAE":"0.31"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1706.08566","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}