View in


L'apprentissage automatique et la production de vaccins

Des experts du CERN apportent leur savoir-faire à Sanofi Pasteur, la division vaccins de Sanofi, multinationale des sciences de la vie


How can machine learning improve vaccine production?

(Image: Robert Gerhardt/ Shutterstock)

Experts from CERN have contributed their know-how in the field of automatic learning to Sanofi Pasteur, the vaccines division of Sanofi, a multinational life sciences company.

A four-day training course focused on topics of particular interest to Sanofi Pasteur was organized with the aim of improving vaccine production. The course focused on the use of ROOT , the high-energy physics data analysis tool, and  TMVA (Toolkit for Multivariate Data Analysis) , a library of algorithms for automatic learning. ROOT has been developed by CERN and various partner institutes and is used by physicists worldwide to analyze data.

The main objective of the course was to see how innovative automated learning techniques could help overcome some problems in vaccine production, which are difficult to resolve with conventional methods. Automatic learning relies on the recognition of patterns among data; the techniques can be used on completely different sets of information. Although CERN's activities have nothing to do with vaccine production, the Laboratory, like the multinational, has to manage many data and multiple variables; automatic learning is therefore an interesting solution.

"This training gave us the opportunity to use and test new methods and understand when they might be useful to us" - Sanofi Pasteur participant.

New possibilities have emerged and several of the teams concerned will test and study the learning tools further. The aim is to use the techniques presented during training to improve the production of vaccines and thus provide more people with access to essential vaccines.

Le cours a été élaboré et donné par Sergei Gleyzer et Lorenzo Moneta, de l'équipe de développement ROOT-TMVA à EP-SFT.  L'accent a également été mis sur les relations avec le CERN et les experts de l'apprentissage automatique, aussi importantes que la formation proprement dite, ce qui ouvre la voie à de nouveaux échanges de connaissances à l'avenir ; le CERN pourra ainsi continuer à contribuer à la création de vaccins.

The course was organized following a conversation between representatives of Sanofi Pasteur and Nick Ziogas of CERN's Knowledge Transfer Group on the tools used at CERN for data analysis. As can be seen, CERN's knowledge and know-how offer many perspectives, but sometimes a simple search on the Internet is not enough to know which ones.