Diogo Barros

Universidade do Minho

10 Abril 2019

Machine Learning in Chemistry: is a machine capable of outsmart a trained chemist?

Each day, 2.5 quintillion bytes of data are generated as a result of the advancements in research which allowed us to maximize the amount of data we can produce. With all this available data, employing machine learning tools became a common practice which has been allowing us the discover of underlying patterns in data that couldn’t be done by other means. Several fields of science had embraced machine learning and conducting interesting breakthroughs from its use and chemistry is no exception, especially in organic synthesis in which novel research has shown that machine learning algorithms can be as good as a skilled chemist (if not better!). But as for anyone who spent hours doing analysis with uncalibrated equipment or watched all 10 Star Wars movies know, quality is more important than quantity. In analytical chemistry, this has been a common problem since the large amount of data collected these days by analytical labs is far from being in a decent shape due to its increased complexity compared to the acquire data from other fields. Nevertheless, interesting discoveries and applications have been made in the last few years. In this talk it will be covered a few of them, ranging from organic synthesis to analytical chemistry, with a special focus in my MSc dissertation.

Projeto POCI-01-0145-FEDER-029147 – PTDC/FIS-PAR/29147/2017 financiado por: OE/FCT, Lisboa 2020, Compete 2020 POCI, Portugal 2020 FEDER