Nicola Marzari, head of the Laboratory of theory and simulation of materials (THEOS) and director of the NCCR MARVEL has been granted an IBM Faculty Award—the first given to an EPFL professor—for the project “Curated Datasets on Demand”.
The IBM Faculty Awards is a competitive worldwide program intended to foster collaboration between researchers at leading universities around the globe and those in IBM research, development and services organizations. The award promotes and recognizes curriculum innovation stimulating growth in disciplines and geographies that are strategic to IBM.
The awarded funding is distributed to work on computational design and the discovery of novel materials, in particular for the exploitation of IBM technologies harvesting both the scientific literature and open-access repositories to train AI models with the purpose of rapidly identifying emerging novel materials. The project developed by Marzari’s group (http://aiida.net) will link open-source solutions with proprietary IBM technologies to flag any material that is the subject of recent studies, according to pre-defined keywords such as superconductivity or thermoelectricity, and proceed automatically to the calculation of the relevant quantities, highlighting any outliers worthy of further study and experimental characterization.
“We have now the capability to calculate key material properties on demand without human intervention – so, doing this for any novel material that appears in the scientific chatter, and flagging anything promising for further study is a natural next step,” Marzari said.
IBM Faculty Awards are cash awards granted annually. The current maximum award to any one recipient is 40,000 USD per year. The awards are not contracts, and no intellectual property rights are required as part of the award. IBM strongly encourages all work to be made available in the public domain.