Molecular Sets (MOSES): A Benchmarking Platform for Molecular Generation Models.

Polykovskiy D, Zhebrak A, Sanchez-Lengeling B, Golovanov S, Tatanov O, Belyaev S, Kurbanov R, Artamonov A, Aladinskiy V, Veselov M, Kadurin A, Johansson S, Chen H, Nikolenko S, Aspuru-Guzik A, Zhavoronkov A. · 2020

Read Molecular Sets (MOSES): A Benchmarking Platform for Molecular Generation Models. by Polykovskiy D, Zhebrak A, Sanchez-Lengeling B, Golovanov S, Tatanov O, Belyaev S, Kurbanov R, Artamonov A, Aladinskiy V, Veselov M, Kadurin A, Johansson S, Chen H, Nikolenko S, Aspuru-Guzik A, Zhavoronkov A. (2020) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/europmc-pmc7775580.txt.

Description

Polykovskiy and co-authors build a benchmark for AI models that invent new molecules. Generative models learn from known compounds and propose fresh structures. Such structures can feed drug screens. But with many models on offer, no shared test existed to rank them fairly. Their platform, MOSES, fills that gap.

MOSES gives standard training and test sets for molecule generation. Its metrics score both quality and diversity of the output. The team ran several known models through the suite. Their scores serve as reference points for new work. The platform and its code are free online for any lab to use and extend.

Catalog record

CREATORPolykovskiy D, Zhebrak A, Sanchez-Lengeling B, Golovanov S, Tatanov O, Belyaev S, Kurbanov R, Artamonov A, Aladinskiy V, Veselov M, Kadurin A, Johansson S, Chen H, Nikolenko S, Aspuru-Guzik A, Zhavoronkov A.
YEAR2020
COLLECTIONPAPERS
FORMATTXT
FILE SIZE45 KB
LICENSECC BY 4.0
RIGHTSCC BY
SOURCEEurope PMC

Source: https://europepmc.org/articles/PMC7775580

Browse the full library · About the Tower of Babel