Philippe Schwaller

CIS – “Get to know your neighbors” Seminar Series

“Accelerating Chemical Synthesis with Transformers”

Prof. Philippe Schwaller, Tenure Track Assistant Professor and Head of LIAC

Monday, Sept. 5, 2022 3:15 – 4:15pm | Hybrid

Add to Google Calendar
Download Calendar file (ics)

or on-site INF 328

 

Note: As this event is reserved to the EPFL community please register with your @epfl.ch email address. 

In organic chemistry, we are currently witnessing a rise in artificial intelligence (AI) approaches, which show great potential for improving molecular designs, facilitating synthesis, and accelerating the discovery of novel molecules. Based on an analogy between written language and organic chemistry, we built linguistics-inspired Transformer models for chemical reaction prediction [1, 2], synthesis planning [3], and the prediction of experimental actions [4,5]. We extended the models to chemical reaction classification and fingerprints [6]. By finding a mapping from discrete reactions to continuous vectors, we enabled efficient chemical reaction space exploration. Intrigued by the remarkable performance of chemical language models, we discovered that the models can capture how atoms rearrange during a reaction, without supervision or human labelling, leading to the development of the open-source atom-mapping tool RXNMapper (http://rxnmapper.ai/) [7]. During my talk, I will provide an overview of the different contributions that are at the base of this digital synthetic chemistry revolution [8].
 
[1]       P. Schwaller et al., ‘Molecular Transformer: A Model for Uncertainty-Calibrated Chemical Reaction Prediction’, ACS Cent. Sci., vol. 5, no. 9, pp. 1572–1583, 2019, doi: 10.1021/acscentsci.9b00576.
[2]       G. Pesciullesi, P. Schwaller, T. Laino, and J.-L. Reymond, ‘Transfer learning enables the molecular transformer to predict regio-and stereoselective reactions on carbohydrates’, Nat. Commun., vol. 11, no. 1, pp. 1–8, 2020.
[3]       P. Schwaller et al., ‘Predicting retrosynthetic pathways using transformer-based models and a hyper-graph exploration strategy’, Chem. Sci., vol. 11, pp. 3316–3325, 2020, doi: 10.1039/C9SC05704H.
[4]       A. C. Vaucher, F. Zipoli, J. Geluykens, V. H. Nair, P. Schwaller, and T. Laino, ‘Automated extraction of chemical synthesis actions from experimental procedures’, Nat. Commun., vol. 11, no. 1, p. 3601, Jul. 2020, doi: 10.1038/s41467-020-17266-6.
[5]       A. C. Vaucher, P. Schwaller, J. Geluykens, V. H. Nair, A. Iuliano, and T. Laino, ‘Inferring experimental procedures from text-based representations of chemical reactions’, Nat. Commun., vol. 12, no. 1, p. 2573, Dec. 2021, doi: 10.1038/s41467-021-22951-1.
[6]       P. Schwaller et al., ‘Mapping the space of chemical reactions using attention-based neural networks’, Nat. Mach. Intell., vol. 3, no. 2, pp. 144–152, Feb. 2021, doi: 10.1038/s42256-020-00284-w.
[7]       P. Schwaller, B. Hoover, J.-L. Reymond, H. Strobelt, and T. Laino, ‘Extraction of organic chemistry grammar from unsupervised learning of chemical reactions’, Sci. Adv., vol. 7, no. 15, p. eabe4166, Apr. 2021, doi: 10.1126/sciadv.abe4166.
[8]       P. Schwaller et al., ‘Machine intelligence for chemical reaction space’, WIREs Comput. Mol. Sci., Mar. 2022, doi: 10.1002/wcms.1604

Philippe Schwaller received a bachelor’s and master’s degree in Materials Science and Engineering from EPFL. While working for IBM Research (2017-2021), Philippe completed an MPhil degree in Physics at the University of Cambridge and a PhD in Chemistry and Molecular Sciences with the Reymond group at the University of Bern. In February 2022, Philippe joined EPFL as a tenure-track assistant professor in the Institute of Chemical Sciences and Engineering. He leads the Laboratory of Artificial Chemical Intelligence (LIAC), which works on AI-accelerated discovery and synthesis of molecules. Philippe is also a core PI of the NCCR Catalysis, a Swiss centre for sustainable chemistry research, education, and innovation.