A team of researchers including NCCR MARVEL’s Michele Ceriotti, EPFL professor and head of the School of Engineering’s Laboratory of Computational Science and Modelling, has received approval from the European Laboratory for Learning and Intelligent Systems (ELLIS) to launch a new ELLIS program, Machine Learning for Molecule Discovery.
Credit: @EPFL
A team of researchers including NCCR MARVEL’s Michele Ceriotti, EPFL professor and head of the School of Engineering’s Laboratory of Computational Science and Modelling, has received approval from the European Laboratory for Learning and Intelligent Systems (ELLIS) to launch a new ELLIS program, Machine Learning for Molecule Discovery.
The program, which is being supported at EPFL by both the NCCR MARVEL and the Centre Européen de Calcul Atomique et Moléculaire (CECAM), will focus on the use of machine learning to accelerate molecular discovery to address pressing societal challenges.
New molecules with specific functions or activities can result in new medicines, secure the world’s food supply via agrochemicals, or enable sustainable energy conversion and storage to counter or mitigate climate change. Discovering new molecules or molecular materials optimized for a particular purpose can take a long time, however, and is highly cost-intensive. Machine learning (ML) methods can accelerate molecular discovery, but domain experts and ML researchers need to work together to ensure that ML has an impact in real world scenarios.
The new ELLIS program, which involves 17 scholars and fellows at institutions including EPFL, ETHZ, the University of Cambridge and MIT, will promote this dialogue in a number of ways. They will look to demonstrate the impact of AI/ML on chemistry, catalysis, drug discovery, energy conversion and molecular modeling in general, as well as advancing fields such as molecular representation learning and property prediction via AI/ML. The program will also facilitate training and education of the next generation of scientists at the interface of chemistry/pharmacology/catalysis and AI/ML through study programs and workshops.
“Molecular and materials discovery is a field in which artificial intelligence can make, and is already making, a real difference,” Ceriotti said. “The ELLIS network has recognized the importance of this topic, and the need for an interdisciplinary team to back this program. We are grateful for the support from CECAM and MARVEL, that stay true to their mission of advancing computational science in chemistry, physics and materials science.”
ELLIS, founded in 2018, is a pan-European AI network of excellence that focuses on fundamental science, technical innovation and societal impact. The network builds on machine learning as the driver for modern AI and aims to secure Europe’s competences and independence in the field through the ELLIS PhD & Postdoc Program; ELLIS Units, including one at EPFL; in addition to the ELLIS programs.
The programs, directed by outstanding European researchers and including leading researchers as Program Fellows, focus on high-impact problem areas. Each has a budget for 2-3 workshops/year to enable meetings between fellows and guests for intensive scientific exchange.
There are currently 14 ELLIS programs covering topics ranging from “Health” to “Theory, Algorithms and Computations of Modern Learning Systems.”