{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,3,28]],"date-time":"2026-03-28T08:09:07Z","timestamp":1774685347500,"version":"3.50.1"},"reference-count":37,"publisher":"MIT Press","issue":"8","license":[{"start":{"date-parts":[[2025,7,8]],"date-time":"2025-07-08T00:00:00Z","timestamp":1751932800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["direct.mit.edu"],"crossmark-restriction":true},"short-container-title":[],"published-print":{"date-parts":[[2025,7,17]]},"abstract":"<jats:title>Abstract<\/jats:title>\n               <jats:p>Generative modeling of crystal structures is significantly challenged by the complexity of input data, which constrains the ability of these models to explore and discover novel crystals. This complexity often confines de novo design methodologies to merely small perturbations of known crystals and hampers the effective application of advanced optimization techniques. One such optimization technique, latent space Bayesian optimization (LSBO), has demonstrated promising results in uncovering novel objects across various domains, especially when combined with variational autoencoders (VAEs). Recognizing LSBO\u2019s potential and the critical need for innovative crystal discovery, we introduce Crystal-LSBO, a de novo design framework for crystals specifically tailored to enhance explorability within LSBO frameworks. Crystal-LSBO employs multiple VAEs, each dedicated to a distinct aspect of crystal structure\u2014lattice, coordinates, and chemical elements\u2014orchestrated by an integrative model that synthesizes these components into a cohesive output. This setup not only streamlines the learning process but also produces explorable latent spaces thanks to the decreased complexity of the learning task for each model, enabling LSBO approaches to operate. Our study pioneers the use of LSBO for de novo crystal design, demonstrating its efficacy through optimization tasks focused mainly on formation energy values. Our results highlight the effectiveness of our methodology, offering a new perspective for de novo crystal discovery.<\/jats:p>","DOI":"10.1162\/neco_a_01767","type":"journal-article","created":{"date-parts":[[2025,7,8]],"date-time":"2025-07-08T13:23:23Z","timestamp":1751981003000},"page":"1505-1527","update-policy":"https:\/\/doi.org\/10.1162\/mitpressjournals.corrections.policy","source":"Crossref","is-referenced-by-count":1,"title":["Crystal-LSBO: Automated Design of De Novo Crystals With Latent Space Bayesian Optimization"],"prefix":"10.1162","volume":"37","author":[{"given":"Onur","family":"Boyar","sequence":"first","affiliation":[{"name":"Department of Mechanical Systems Engineering, Nagoya University, Furo-cho, Chikusa-ku, Nagoya, Aichi 4648603, Japan"},{"name":"RIKEN Center for Advanced Intelligence Project, Furo-cho, Chikusa-ku, Nagoya, Aichi 4648603, Japan boyar.onur.nagoyaml@gmail.com"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yanheng","family":"Gu","sequence":"additional","affiliation":[{"name":"Department of Mechanical Systems Engineering, Nagoya University, Furo-cho, Chikusa-ku, Nagoya, Aichi 4648603, Japan gu.yanheng.nagoyaml@gmail.com"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yuji","family":"Tanaka","sequence":"additional","affiliation":[{"name":"Department of Mechanical Systems Engineering, Nagoya University, Furo-cho, Chikusa-ku, Nagoya, Aichi 4648603, Japan tanaka.yuji.nagoyaml@gmail.com"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Shunsuke","family":"Tonogai","sequence":"additional","affiliation":[{"name":"DENSO, Kariya, Aichi 4488661, Japan shunsuke.tonogai.j3y@jp.denso.com"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Tomoya","family":"Itakura","sequence":"additional","affiliation":[{"name":"DENSO, Kariya, Aichi 4488661, Japan tomoya.itakura.j8w@jp.denso.com"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Ichiro","family":"Takeuchi","sequence":"additional","affiliation":[{"name":"Department of Mechanical Systems Engineering, Nagoya University, Furo-cho, Chikusa-ku, Nagoya, Aichi 4648603, Japan"},{"name":"RIKEN Center for Advanced Intelligence Project, Furo-cho, Chikusa-ku, Nagoya, Aichi 4648603, Japan takeuchi.ichiro.n6@f.mail.nagoya-u.ac.jp"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"281","published-online":{"date-parts":[[2025,7,17]]},"reference":[{"key":"2025072212153403800_bib1","author":"Abenza","year":"2023","journal-title":"Data augmentation for battery materials using lattice scaling"},{"key":"2025072212153403800_bib1a","volume":"A.","author":"Arnold","year":"2016","journal-title":"International tables for crystallography"},{"issue":"11","key":"2025072212153403800_bib2","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1162\/neco_a_01708","article-title":"Latent space Bayesian optimization with latent data augmentation for enhanced exploration","volume":"36","author":"Boyar","year":"2024","journal-title":"Neural Computation"},{"issue":"11","key":"2025072212153403800_bib3","doi-asserted-by":"publisher","first-page":"718","DOI":"10.1038\/s43588-022-00349-3","article-title":"A universal graph deep learning interatomic potential for the periodic table","volume":"2","author":"Chen","year":"2022","journal-title":"Nature Computational Science"},{"key":"2025072212153403800_bib4","doi-asserted-by":"crossref","first-page":"785","DOI":"10.1145\/2939672.2939785","article-title":"XGBoost: A scalable tree boosting system","volume-title":"Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining","author":"Chen","year":"2016"},{"key":"2025072212153403800_bib5","author":"Chiba","year":"2022","journal-title":"Neural structure fields with application to crystal structure autoencoders"},{"issue":"10","key":"2025072212153403800_bib6","doi-asserted-by":"publisher","first-page":"4518","DOI":"10.1021\/acs.jcim.0c00464","article-title":"3-D inorganic crystal structure generation and property prediction via representation learning","volume":"60","author":"Court","year":"2020","journal-title":"Journal of Chemical Information and Modeling"},{"key":"2025072212153403800_bib7","first-page":"493","article-title":"High-dimensional Bayesian optimization with sparse axis-aligned subspaces","volume-title":"Proceedings of the Thirty-Seventh Conference on Uncertainty in Artificial Intelligence","author":"Eriksson","year":"2021"},{"issue":"2","key":"2025072212153403800_bib8","doi-asserted-by":"crossref","first-page":"268","DOI":"10.1021\/acscentsci.7b00572","article-title":"Automatic chemical design using a data-driven continuous representation of molecules","volume":"4","author":"G\u00f3mez-Bombarelli","year":"2018","journal-title":"ACS Central Science"},{"key":"2025072212153403800_bib9","doi-asserted-by":"publisher","first-page":"577","DOI":"10.1039\/C9SC04026A","article-title":"Constrained Bayesian optimization for automatic chemical design using variational autoencoders","volume":"11","author":"Griffiths","year":"2020","journal-title":"Chemical Science"},{"key":"2025072212153403800_bib10","author":"Grosnit","year":"2021","journal-title":"High-dimensional Bayesian optimisation with variational autoencoders and deep metric learning"},{"key":"2025072212153403800_bib11","article-title":"Beta-VAE: Learning basic visual concepts with a constrained variational framework","volume-title":"Proceedings of the International Conference on Learning Representations","author":"Higgins","year":"2016"},{"key":"2025072212153403800_bib12","article-title":"RMSProp: Divide the gradient by a running average of its recent magnitude","author":"Hinton","year":"2012","journal-title":"COURSERA: Neural Networks for Machine Learning"},{"key":"2025072212153403800_bib13","author":"Hoffmann","year":"2019","journal-title":"Data-driven approach to encoding and decoding 3-D crystal structures"},{"key":"2025072212153403800_bib14","first-page":"20793","article-title":"Vanilla Bayesian optimization performs great in high dimensions","volume-title":"Proceedings of the 41st International Conference on Machine Learning","author":"Hvarfner","year":"2024"},{"key":"2025072212153403800_bib15","first-page":"448","article-title":"Batch normalization: Accelerating deep network training by reducing internal covariate shift","volume-title":"Proceedings of the 32nd International Conference on Machine Learning","author":"Ioffe","year":"2015"},{"key":"2025072212153403800_bib16","doi-asserted-by":"publisher","DOI":"10.1063\/1.4812323","article-title":"Commentary: The Materials Project: A materials genome approach to accelerating materials innovation","volume":"1","author":"Jain","year":"2013","journal-title":"APL Materials"},{"key":"2025072212153403800_bib17","first-page":"2323","article-title":"Junction tree variational autoencoder for molecular graph generation","volume-title":"Proceedings of the International Conference on Machine Learning","author":"Jin","year":"2018"},{"issue":"8","key":"2025072212153403800_bib18","doi-asserted-by":"publisher","first-page":"1412","DOI":"10.1021\/acscentsci.0c00426","article-title":"Generative adversarial networks for crystal structure prediction","volume":"6","author":"Kim","year":"2020","journal-title":"ACS Central Science"},{"key":"2025072212153403800_bib19","article-title":"Auto-encoding variational Bayes","volume-title":"Proceedings of the 2nd International Conference on Learning Representations","author":"Kingma","year":"2014"},{"issue":"4","key":"2025072212153403800_bib20","doi-asserted-by":"publisher","DOI":"10.1088\/2632-2153\/aba947","article-title":"Self- referencing embedded strings (SELFIES): A 100% robust molecular string representation","volume":"1","author":"Krenn","year":"2020","journal-title":"Machine Learning: Science and Technology"},{"key":"2025072212153403800_bib21","first-page":"34505","article-title":"Local latent space Bayesian optimization over structured Inputs","volume-title":"Advances in neural information processing systems","author":"Maus","year":"2022"},{"key":"2025072212153403800_bib22","author":"McInnes","year":"2018","journal-title":"UMAP: Uniform manifold approximation and projection for dimension reduction"},{"issue":"5","key":"2025072212153403800_bib23","doi-asserted-by":"publisher","first-page":"1370","DOI":"10.1016\/j.matt.2019.08.017","article-title":"Inverse design of solid-state materials via a continuous representation","volume":"1","author":"Noh","year":"2019","journal-title":"Matter"},{"key":"2025072212153403800_bib24","doi-asserted-by":"publisher","first-page":"314","DOI":"10.1016\/j.commatsci.2012.10.028","article-title":"Python materials genomics (pymatgen): A robust, open-source Python library for materials analysis","volume":"68","author":"Ong","year":"2013","journal-title":"Computational Materials Science"},{"key":"2025072212153403800_bib25","article-title":"PyTorch: An imperative style, high-performance deep learning library","volume-title":"Advances in neural information processing systems","author":"Paszke","year":"2019"},{"key":"2025072212153403800_bib26","author":"Ren","year":"2020","journal-title":"Inverse design of crystals using generalized invertible crystallographic representation"},{"issue":"1","key":"2025072212153403800_bib27","doi-asserted-by":"publisher","first-page":"314","DOI":"10.1016\/j.matt.2021.11.032","article-title":"An invertible crystallographic representation for general inverse design of inorganic crystals with targeted properties","volume":"5","author":"Ren","year":"2022","journal-title":"Matter"},{"key":"2025072212153403800_bib28","article-title":"Sample-efficient optimization in the latent space of deep generative models via weighted retraining","volume-title":"Advances in neural information processing systems","author":"Tripp","year":"2020"},{"key":"2025072212153403800_bib29","article-title":"High-dimensional Bayesian optimization with invariance","volume-title":"Proceedings of the ICML Workshop on Adaptive Experimental Design and Active Learning","author":"Verma","year":"2022"},{"issue":"3","key":"2025072212153403800_bib30","first-page":"33","article-title":"One method of multiextremum optimization","volume":"5","author":"Vydunas","year":"1971","journal-title":"Avtomatika i Vychislitel\u2019naya Tekhnika. Automatic Control and Computer Sciences"},{"key":"2025072212153403800_bib31","doi-asserted-by":"publisher","first-page":"361","DOI":"10.1613\/jair.4806","article-title":"Bayesian optimization in a billion dimensions via random embeddings","volume":"55","author":"Wang","year":"2016","journal-title":"Journal of Artificial Intelligence Research"},{"issue":"1","key":"2025072212153403800_bib32","doi-asserted-by":"publisher","first-page":"31","DOI":"10.1021\/ci00057a005","article-title":"SMILES, a chemical language and information system. 1. Introduction to methodology and encoding rules","volume":"28","author":"Weininger","year":"1988","journal-title":"Journal of Chemical Information and Computer Sciences"},{"key":"2025072212153403800_bib33","author":"Xie","year":"2021","journal-title":"Crystal diffusion Variational autoencoder for periodic material generation"},{"key":"2025072212153403800_bib34","author":"Zeni","year":"2023","journal-title":"MatterGen: A generative model for inorganic materials design"},{"issue":"20","key":"2025072212153403800_bib35","doi-asserted-by":"crossref","DOI":"10.1002\/advs.202100566","article-title":"High-throughput discovery of novel cubic crystal materials using deep generative neural networks","volume":"8","author":"Zhao","year":"2021","journal-title":"Advanced Science"},{"issue":"1","key":"2025072212153403800_bib36","article-title":"Physics guided deep learning for generative design of crystal materials with symmetry constraints","volume":"9","author":"Zhao","year":"2023","journal-title":"npj Computational Materials"}],"container-title":["Neural Computation"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/direct.mit.edu\/neco\/article-pdf\/37\/8\/1505\/2531744\/neco_a_01767.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/direct.mit.edu\/neco\/article-pdf\/37\/8\/1505\/2531744\/neco_a_01767.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,7,22]],"date-time":"2025-07-22T16:15:41Z","timestamp":1753200941000},"score":1,"resource":{"primary":{"URL":"https:\/\/direct.mit.edu\/neco\/article\/37\/8\/1505\/131386\/Crystal-LSBO-Automated-Design-of-De-Novo-Crystals"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,7,17]]},"references-count":37,"journal-issue":{"issue":"8","published-online":{"date-parts":[[2025,7,17]]},"published-print":{"date-parts":[[2025,7,17]]}},"URL":"https:\/\/doi.org\/10.1162\/neco_a_01767","relation":{},"ISSN":["0899-7667","1530-888X"],"issn-type":[{"value":"0899-7667","type":"print"},{"value":"1530-888X","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2025,8]]},"published":{"date-parts":[[2025,7,17]]}}}