As the pandemic continues its devastating effects and we race against time to release a vaccine in the public domain, the production questions haunt.
Even if we get an effective vaccine will we be able to produce enough to distribute among people? Is the process fast enough? The answer to all these questions lies in Machine Learning. An old article published in the CERN website highlighting CERN scientists speaking on the use of machine learning in vaccine and drug development comes to the forefront.
A Machine Learning Vaccine Production Course
- A team of experts from CERN shared their expertise on machine learning with Sanofi Pasteur, the vaccines business unit of Sanofi, a global life sciences company.
- This was achieved by means of a four-day training course tailored to address topics specifically of interest to Sanofi-Pasteur, with the aim of improving vaccine production.
- The course was built around ROOT, the data analysis framework used to analyse HEP data, and the Toolkit for Multivariate Data Analysis (TMVA), a library of associated machine learning algorithms.
- ROOT was developed by CERN and various collaborating institutes and is used by physicists around the globe to analyse data.
How machine learning can help?
The main objective of the course was to apply novel machine learning techniques to various vaccine production challenges that had proven hard to solve using conventional methods.
Machine learning is all about finding patterns in data and the techniques can be exploited across completely different sets of information.
How physicists in CERN can help?
Despite the fact that CERN has nothing to do with vaccine production, both organisations have plenty of data and many variables, making machine learning valuable.
“This training course gave us the opportunity to use and test new methods and understand in which cases they could be useful for us,” said a participant from Sanofi Pasteur.
New opportunities came to light and several of the teams involved will test and explore machine learning tools further.
The aim is that the techniques discussed during the training course could be deployed to improve vaccine production and consequently help even more people to access vital vaccines.
The course was prepared and delivered by Sergei Gleyzer and Lorenzo Moneta from the ROOT-TMVA development team in EP-SFT. It was emphasised that the relationship with CERN and the machine learning experts was just as valuable as the training course itself, which leaves the possibility for further knowledge exchange in the future, allowing CERN to continue to aid the creation of vaccines.
How was it organized?
The training course was organised following a face-to-face conversation between representatives of Sanofi Pasteur and Nick Ziogas from CERN’s Knowledge Transfer group, about tools used at CERN for data analysis. As this illustrates, there are many opportunities to learn from CERN’s knowledge and expertise, but sometimes it takes more than an internet search to identify them.