Deep Learning Spectroscopy: Neural Networks for Molecular Excitation Spectra

dc.contributorAalto-yliopistofi
dc.contributorAalto Universityen
dc.contributor.authorGhosh, Kunalen_US
dc.contributor.authorStuke, Annikaen_US
dc.contributor.authorTodorović, Milicaen_US
dc.contributor.authorJørgensen, Peter Bjørnen_US
dc.contributor.authorSchmidt, Mikkel N.en_US
dc.contributor.authorVehtari, Akien_US
dc.contributor.authorRinke, Patricken_US
dc.contributor.departmentDepartment of Applied Physicsen
dc.contributor.departmentDepartment of Computer Scienceen
dc.contributor.groupauthorComputational Electronic Structure Theoryen
dc.contributor.groupauthorProbabilistic Machine Learningen
dc.contributor.groupauthorProfessorship Vehtari Akien
dc.contributor.groupauthorHelsinki Institute for Information Technology (HIIT)en
dc.contributor.organizationDanmarks Tekniske Universiteten_US
dc.date.accessioned2019-02-25T08:50:46Z
dc.date.available2019-02-25T08:50:46Z
dc.date.issued2019-05-03en_US
dc.description| openaire: EC/H2020/676580/EU//NoMaD
dc.description.abstractDeep learning methods for the prediction of molecular excitation spectra are presented. For the example of the electronic density of states of 132k organic molecules, three different neural network architectures: multilayer perceptron (MLP), convolutional neural network (CNN), and deep tensor neural network (DTNN) are trained and assessed. The inputs for the neural networks are the coordinates and charges of the constituent atoms of each molecule. Already, the MLP is able to learn spectra, but the root mean square error (RMSE) is still as high as 0.3 eV. The learning quality improves significantly for the CNN (RMSE = 0.23 eV) and reaches its best performance for the DTNN (RMSE = 0.19 eV). Both CNN and DTNN capture even small nuances in the spectral shape. In a showcase application of this method, the structures of 10k previously unseen organic molecules are scanned and instant spectra predictions are obtained to identify molecules for potential applications.en
dc.description.versionPeer revieweden
dc.format.extent1-7
dc.format.mimetypeapplication/pdfen_US
dc.identifier.citationGhosh, K, Stuke, A, Todorović, M, Jørgensen, P B, Schmidt, M N, Vehtari, A & Rinke, P 2019, ' Deep Learning Spectroscopy : Neural Networks for Molecular Excitation Spectra ', Advanced Science, vol. 6, no. 9, 1801367, pp. 1-7 . https://doi.org/10.1002/advs.201801367en
dc.identifier.doi10.1002/advs.201801367en_US
dc.identifier.issn2198-3844
dc.identifier.otherPURE UUID: 9ebd485e-533e-4479-8336-1adbe3715866en_US
dc.identifier.otherPURE ITEMURL: https://research.aalto.fi/en/publications/9ebd485e-533e-4479-8336-1adbe3715866en_US
dc.identifier.otherPURE LINK: http://www.scopus.com/inward/record.url?scp=85060767751&partnerID=8YFLogxKen_US
dc.identifier.otherPURE FILEURL: https://research.aalto.fi/files/31971829/Ghosh_et_al_2019_Advanced_Science.pdfen_US
dc.identifier.urihttps://aaltodoc.aalto.fi/handle/123456789/36841
dc.identifier.urnURN:NBN:fi:aalto-201902251998
dc.language.isoenen
dc.publisherWiley - VCH
dc.relationinfo:eu-repo/grantAgreement/EC/H2020/676580/EU//NoMaDen_US
dc.relation.ispartofseriesAdvanced Scienceen
dc.rightsopenAccessen
dc.subject.keywordartificial intelligenceen_US
dc.subject.keywordDFT calculationsen_US
dc.subject.keywordexcitation spectraen_US
dc.subject.keywordneural networksen_US
dc.subject.keywordorganic moleculesen_US
dc.titleDeep Learning Spectroscopy: Neural Networks for Molecular Excitation Spectraen
dc.typeA1 Alkuperäisartikkeli tieteellisessä aikakauslehdessäfi
dc.type.versionpublishedVersion
Files