ICME 2023: Mat Data & Platforms: I
Program Organizers: Charles Ward, AFRL/RXM; Heather Murdoch, U.S. Army Research Laboratory
Monday 9:50 AM
May 22, 2023
Room: Caribbean IV
Location: Caribe Royale
Session Chair: Fatih Sen, Novelis
9:50 AM Invited
Materials Data & Informatics: Curation, Frameworks, Access, and Potential for Discovery and Design: L. Catherine Brinson1; 1Duke University
With the advent of the materials genome initiative (MGI) in the United States and a similar focus on materials data around the world, numerous materials data resources and associated vocabularies, tools, and repositories have been developed. While the majority of these systems focus on slices of computational data with an emphasis on crystallographic materials, platforms for organic materials and their composites, especially those incorporating experimental data, have been quite limited. We will discuss the unique aspects of tackling data assembly and informatics associated with experimental organic materials data, with focus on our experiences creating an open-source data resource, NanoMine, part of MaterialsMine. Our goal has been to curate, annotate and store widely varying experimental data on polymer nanocomposites (polymers doped with nanofiller) and providing access to characterization and analysis tools with the long-term objective of promoting facile nanocomposite design. The challenges and promises associated with data curation, ontology and vocabulary development, standardization and interoperability, and data visualization and analysis tools will be discussed. Several case studies will be presented, including use of natural language processing for archival data curation, coupling of experimental and computational data for materials design, and development of machine learning tools for rapid property screening and inference. Overall, we focus on the promise of this new approach to tackle materials design principles for the complex, high dimensional problems inherent in the multi-phase polymer space.
10:20 AM
FAIR Data in PMD:
Development of MSE Mid-level and Standard-compliant Application Ontologies: Markus Schilling1; Bernd Bayerlein1; Philipp von Hartrott2; Jörg Waitelonis3; Henk Birkholz4; Jannis Grundmann5; Pedro Portella2; Birgit Skrotzki1; 1Federal Ministry of Materials Research and Testing; 2Fraunhofer Institute for Mechanics of Materials; 3Leibniz Institute for Information Infrastructure; 4Leibniz-IWT Institut für Werkstofforientierte Technologien ; 5Leibniz-IWT Institut für Werkstofforientierte Technologien
The efforts taken within the project ‘platform MaterialDigital’ (PMD, materialdigital.de) to store FAIR data in accordance with a standard-compliant ontological representation (‘application ontology’) of a tensile test of metals at room temperature (ISO 6892-1:2019-11) will be presented. This includes the path from developing an ontology in accordance with the respective standard, converting ordinary data obtained from standard tests into the interoperable RDF format, up to connecting the ontology and data. The semantic connection of the ontology and data leads to interoperability and an enhanced ability of querying. For further reusability of data and knowledge semantically stored, the PMD core ontology (PMDco) was developed, which is a mid-level ontology in the field of MSE. The semantic connection of the tensile test application ontology to the PMDco is also presented. Moreover, Ontopanel, a tool for domain experts facilitating visual ontology development and mapping for FAIR data sharing in MSE, is introduced briefly.
10:40 AM Cancelled
Automatic Deducing the New Materials Knowledge within the OWL Framework: Evgeny Blokhin1; 1Tilde Materials Informatics
The computer ontologies (“ashes from the fire of human thinking”) can be practically thought as the advanced databases. Their OWL standard was developed with the aim to mathematically guarantee, that the logical reasoning succeeds in finite time. In this work I benchmark the modern Python tools to work with several established nowadays OWL materials ontologies, such as EMMO and PMD. I also present our own in-house effort called MPDS. The highly inter-linked MPDS data graph consists of approximately 5M nodes and 150M edges. All the assertions were taken from the world’s published literature in materials science (about 0.5M publications since 1891), as curated by the Pauling Files project, see www.mpds.io. Also the MPDS data graph can be greatly expanded with our in-house ab initio simulations data, which will be produced high-throughput in the cloud fully automatically, controlled with the ontology reasoning engine.
11:00 AM
NIST Interatomic Potentials Repository: Discovering, Evaluating and Comparing Interatomic Potentials: Lucas Hale1; 1National Institute of Standards and Technology
Classical atomistic calculations provide a means of exploring the important linkages between atomic structures and larger scale dynamic properties and behaviors. However, the calculation predictions are strongly dependent on the choice of interatomic potential used. The NIST Interatomic Potentials Repository contains citation listings and parameter files for known potentials, as well as a large collection of computed crystalline and crystal defect properties specific to each potential. This not only makes it easy to discover existing potentials, but helps users select which potentials are best suited for their interests. All calculation methods, underlying tools, and high throughput capabilities are available as open-source Python code. The entire calculation framework is designed to be extensible and accessible to users at all levels of interaction.
11:20 AM
Materials Commons and FAIR Data: Glenn Tarcea1; John Allison1; Brian Puchala1; Tracy Berman1; 1University of Michigan
The Materials Commons is a data repository for the materials sciences. We describe how Materials Commons conforms to FAIR Data and discuss some of the challenges and needs to make data reusable. The FAIR Data principles of Findable, Accessible, Interoperable and Reusable provide a framework for data reuse, but there is still a large gap to achieving true reusability and understandability. Materials Commons has been extending the ideas of FAIR to help close this gap. We discuss the need for cultural and funding agency changes in order to achieve the larger goal of reuse. We look into how technology, as well as incentives from research funding agencies, can help make published data more reusable.