BIST Colloquium Series 2021-22
by Teodoro Laino, IBM Research Europe, Switzerland
Are Language Models better than Physical based Models for Chemistry (and more)?
Natural language processing models in chemistry have emerged as one of the most effective approaches for capturing human knowledge and modelling creativity in organic chemistry. Its application in machine learning tasks demonstrated high quality and ease of use in problems such as predicting chemical reactions [1-2], retrosynthetic routes , digitizing chemical literature , predicting detailed experimental procedures , designing new fingerprints  and yield predictions . In this presentation, I will discuss the impact of language models in chemistry by describing first the implementation of the first cloud-based AI-driven autonomous laboratory .
Finally, I will focus on the most recent applications of language models to address the problem of characterizing unknown enzymes, the development of human-in-the-loop schemes for retrosynthetic strategies or promoting sustainability and green chemistry strategies with ad-hoc AI models.
- IBM Research Europe, Chem. Sci., 2018, 9, 6091-6098
- IBM Research Europe, ACS Cent. Sci. 2019, 5, 9, 1572-1583
- IBM Research Europe, Chem. Sci., 2020, 11, 3316-3325
- IBM Research Europe, Nat. Comm., 2020, 11, 3601
- IBM Research Europe, Nat. Comm., 2021, 12, 2573
- IBM Research Europe, Nat. Mach. Intel., 2021, 3, 144–152
- IBM Research Europe, Mach. Learn.: Sci. Technol., 2021, 2, 015016
I received my degree in theoretical chemistry in 2001 (University of Pisa and Scuola Normale Superiore di Pisa) and the doctorate in 2006 in computational chemistry at the Scuola Normale Superiore di Pisa, Italy. My doctoral thesis, entitled “Multi-Grid QM/ MM Approaches in ab initio Molecular Dynamics” was supervised by Prof. Dr. Michele Parrinello. From 2006 to 2008, I worked as a post-doctoral researcher in the research group of Prof. Dr. Jürg Hutter at the University of Zurich, where I developed algorithms for ab initio and classical molecular dynamics simulations. Since 2008, I have been working in the department of Cognitive Computing and Industry Solutions at the IBM Research – Zurich Laboratory (ZRL).
The focus of my research is on complex material simulations for industrial-related problems (energy storage, life sciences and nano-electronics). More recently, I am interested by the application of machine learning/artificial intelligence technologies to chemistry and materials science problems with the purpose of developing customized solutions as (for example) IBM RXN for chemistry.
This colloquium is part of the BIST Master of Multidisciplinary Research in Experimental Sciences curriculum but is also open and free for anyone interested in participating. If you want to assist please send an email to email@example.com