{"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/hierarchical-modeling-of-molecular-energies","title":"Hierarchical modeling of molecular energies using a deep neural network","arxiv_id":"1710.00017","date":"2017-09-29","proceeding":null,"authors":["Nicholas Lubbers","Justin S. Smith","Kipton Barros"],"abstract":"We introduce the Hierarchically Interacting Particle Neural Network (HIP-NN)\nto model molecular properties from datasets of quantum calculations. Inspired\nby a many-body expansion, HIP-NN decomposes properties, such as energy, as a\nsum over hierarchical terms. These terms are generated from a neural network--a\ncomposition of many nonlinear transformations--acting on a representation of\nthe molecule. HIP-NN achieves state-of-the-art performance on a dataset of 131k\nground state organic molecules, and predicts energies with 0.26 kcal/mol mean\nabsolute error. With minimal tuning, our model is also competitive on a dataset\nof molecular dynamics trajectories. In addition to enabling accurate energy\npredictions, the hierarchical structure of HIP-NN helps to identify regions of\nmodel uncertainty.","url_abs":"http://arxiv.org/abs/1710.00017v1","url_pdf":"http://arxiv.org/pdf/1710.00017v1.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":[],"tasks":[{"task_slug":"drug-discovery","task_name":"Drug Discovery"},{"task_slug":"formation-energy","task_name":"Formation Energy"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/formation-energy-on-qm9","task":"Formation Energy","dataset":"QM9","model":"HIP-NN","rank_in_archive_order":13,"of":18,"metrics":{"MAE":"0.256"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/1710.00017","atlas_url":"https://app.syntology.ai/?focus=1710.00017","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}