PuSH - Publikationsserver des Helmholtz Zentrums München

Export: TXT Text RIS Endnote (RIS) BIB BibTeX
Karpov, P. ; Godin, G.* ; Tetko, I.V.

A transformer model for retrosynthesis.

Lect. Notes Comput. Sc. 11731 LNCS, 817-830 (2019)
Verlagsversion DOI
Open Access Hybrid
Creative Commons Lizenzvertrag
We describe a Transformer model for a retrosynthetic reaction prediction task. The model is trained on 45 033 experimental reaction examples extracted from USA patents. It can successfully predict the reactants set for 42.7% of cases on the external test set. During the training procedure, we applied different learning rate schedules and snapshot learning. These techniques can prevent overfitting and thus can be a reason to get rid of internal validation dataset that is advantageous for deep models with millions of parameters. We thoroughly investigated different approaches to train Transformer models and found that snapshot learning with averaging weights on learning rates minima works best. While decoding the model output probabilities there is a strong influence of the temperature that improves at $$\text {T}=1.3$$ the accuracy of models up to 1–2%.
Impact Factor
Scopus SNIP
Scopus
Cited By
Altmetric
0.000
0.713
49
Tags
Anmerkungen
Besondere Publikation
Auf Hompepage verbergern

Zusatzinfos bearbeiten
Eigene Tags bearbeiten
Privat
Eigene Anmerkung bearbeiten
Privat
Auf Publikationslisten für
Homepage nicht anzeigen
Als besondere Publikation
markieren
Publikationstyp Artikel: Journalartikel
Dokumenttyp Wissenschaftlicher Artikel
Schlagwörter Character-based Models ; Computer Aided Synthesis Planning ; Retrosynthesis Prediction ; Transformer
Sprache
Veröffentlichungsjahr 2019
HGF-Berichtsjahr 2019
ISSN (print) / ISBN 0302-9743
e-ISSN 1611-3349
Zeitschrift Lecture Notes in Computer Science
Quellenangaben Band: 11731 LNCS, Heft: , Seiten: 817-830 Artikelnummer: , Supplement: ,
Verlag Springer
Verlagsort Berlin [u.a.]
Institut(e) Institute of Structural Biology (STB)
POF Topic(s) 30203 - Molecular Targets and Therapies
Forschungsfeld(er) Enabling and Novel Technologies
PSP-Element(e) G-503000-001
DOI 10.1007/978-3-030-30493-5_78
Scopus ID 85072983479
Erfassungsdatum 2019-10-16
[➜Korrektur melden]