NOMAD establishes new, fourth paradigm in computational materials science
Claudia Draxl and Matthias Scheffler were invited to write an article about the NOMAD Laboratory CoE and their view about data-driven materials science for the MRS Bulletin, a journal of the US Materials Research Society. The Topical Issue on Data-Centric Science and Materials Innovation, whose guest Editors are Isao Tanaka (Kyoto University), Krishna Rajan (University of Buffalo), and Chris Wolverton (Northwestern University), will appear later this year (preprint at http://arxiv.org/abs/1805.05039).
Their contribution “NOMAD: The FAIR Concept for Big-Data-Driven Materials Science” elaborates on the FAIR principles (Findable, Accessible, Interoperable, Re-usable) that have been developed independently and parallel to NOMAD. They also explain the four paradigms of materials science and engineering (Figure 1).
Figure 1. Development of paradigms (new modes of thought) of materials science and engineering
The vision of NOMAD was, and is, to establish the fourth paradigm in computational materials science. In this paradigm, it is recognized that Big-Data contain correlations, reflected in terms of structure and patterns in the data that are not visible in small data sets, and that many properties of materials cannot be described by a closed mathematical form as they are determined by several intricate, multi-level theoretical concepts.
The paper elaborates on all five elements of the NOMAD CoE (Repository, Archive, Encyclopedia, Big-Data Analytics and Visualization), and shows the recent upload statistics, which reflects how NOMAD has been accepted and is being used by the community. NOMAD supports all important codes in computational materials science (Figure 2).
Figure 2. Number of uploaded open-access total-energy calculations in the NOMAD Repository and Archive as of March 15, 2018. Codes with more than 80 uploads are shown.
To date, 40 codes from electronic-structure theory and force fields are supported by NOMAD. As noted above, the NOMAD concepts are 100% FAIR. The total number of open-access total-energy calculations at the NOMAD Repository and Archive is more than 50 million, corresponding to billions of CPU-core hours.