Design and Experimental Validation of a Photocatalyst Recommender Based on a Large Language Model

Simple item record

Full item record

dc.contributor.authorMillward, Francis
dc.contributor.authorKulczykowski, Michał
dc.contributor.authorBadland-Shaw, Jay
dc.contributor.authorSzymkuc, Sara
dc.contributor.authorSuraksha, Rajan
dc.contributor.authorSrivastawa, Aniket Kumar
dc.contributor.authorManet, Violaine
dc.contributor.authorGriffin, Máire
dc.contributor.authorBryden, Megan
dc.contributor.authorComerford, Thomas
dc.contributor.authorHämmerling, Lea
dc.contributor.authorMariko, Aminata
dc.contributor.authorGrzybowski, Bartosz A.
dc.contributor.authorZysman-Colman, Eli
dc.contributor.organizationOrganic Semiconductor Centre, EaStCHEM School of Chemistry, University of St Andrews, UK
dc.contributor.organizationAllchemy, Highland, Indiana, USA
dc.contributor.organizationInstitute of Organic Chemistry, Polish Academy of Sciences
dc.contributor.organizationCenter for Algorithmic and Robotized Synthesis (CARS) of Korea’s Institute for Basic Science (IBS) and Department of Chemistry, Ulsan National Institute of Science and Technology, South Korea
dc.date.accessioned2026-01-30T17:31:05Z
dc.date.available2026-01-30T17:31:05Z
dc.date.issued2026
dc.date.submitted2026-01-30T08:16:19Zen
dc.description.abstractUtilizing an extensive library of literature on photocatalytic transformations, we disclose the development of a machine learning (ML) model for the recommendation of photocatalysts most suitable for reactions of interest. The model is trained on > 36 000 such literature examples and uses an architecture inspired by the Bidirectional Encoder Representations from Transformer (BERT) large language model. Under cross-validation, it can suggest the “correct” photocatalysts with ∼90% accuracy. When experimentally tested on five out-of-box reactions, this algorithm consistently suggested photocatalysts that gave yields competitive to those chosen by human researchers and frequently suggested alternative photocatalysts that are potentially more appealing than the originally selected photocatalyst. Altogether, this platform serves as a valuable tool for researchers undertaking reaction optimization programs. The model is free to use at https://photocatals.grzybowskigroup.pl/predict/. Choosing a photocatalyst for a given reaction can be challenging due to complex mechanisms and multiple parameters that govern the outcome of a photocatalyzed reaction. Herein, we disclose a machine learning (ML) model that can suggest catalysts for a given reaction using an online portal. The model was experimentally validated against five photocatalysis reactions, in all cases suggesting productive photocatalysts. This model serves as a valuable tool for researchers optimizing photocatalysis reactions.en
dc.description.sponsorshipInstitute for Basic Science, Korea (project code IBS-R020-D1); Leverhulme Trust (RPG-2023–110); Engineering and Physical Sciences Research Council (EP/W007517/1, EP/W015137/1, EP/Z535291/1, and EP/W522259/1); European Commission (PhotoReAct ITN: 956324).
dc.identifier.citationAngew. Chem. Int. Ed. 2026, 65, e14544 (1 of 10) // https://doi.org/10.1002/anie.202514544
dc.identifier.doi10.1002/anie.202514544
dc.identifier.issn1433-7851
dc.identifier.issn1521-3773
dc.identifier.urihttps://open.icm.edu.pl/handle/123456789/26463
dc.language.isoen
dc.publisherWiley-VCH GmbH
dc.relation.ispartofseries65
dc.rightsUznanie autorstwa 4.0 Międzynarodoween
dc.rights.urihttp://creativecommons.org/licenses/by/4.0/
dc.sourceAngewandte Chemie International Edition
dc.subjectlarge language modelsen
dc.subjectmachine learningen
dc.subjectphotocatalysisen
dc.titleDesign and Experimental Validation of a Photocatalyst Recommender Based on a Large Language Modelen
dc.typearticle
dc.type.versionpublishedVersion
person.identifier.orcidGrzybowski, Bartosz A. [0000-0001-6613-4261]
person.identifier.orcidMillward, Francis [0000-0002-6342-1326]
person.identifier.orcidSuraksha, Rajan [0009-0002-8426-7038]
person.identifier.orcidGriffin, Máire [0000-0003-4177-0864]
person.identifier.orcidBryden, Megan [0000-0003-2542-4427]
person.identifier.orcidComerford, Thomas [0000-0002-5966-5686]
person.identifier.orcidMariko, Aminata [0000-0003-4441-1816]
person.identifier.orcidZysman-Colman, Eli [0000-0001-7183-6022]
Files for this record
Original bundle
Now showing 1 - 1 of 1
Name: Angew Chem Int Ed - 2025 - Millward - Design and Experimental Validation of a Photocatalyst Recommender Based on a Large.pdf
Size: 1.85 MB
Format: Adobe Portable Document Format
Description:
Download
License files
Name: license_rdf
Size: 1019 B
Format: RDF serialized in XML
Description:
Download
Belongs to collection
Artykuły IChO PAN