Advancing Open Science for Societal Impact (Open Data and Open Education)

Dr. Nong Artrith's research group aligns with Utrecht University's mission to accelerate and enhance the realization of research results and their societal impact through the principles of Open Science https://www.uu.nl/en/research/open-science . 

2024 Dr. Artrith wins Sebastian Haben award for Open Science.

Nong Artrith is a computational materials scientist investigating fundamental processes for clean energy storage and conversion using quantum-mechanics-based atomic-scale simulations and artificial intelligence (AI). For her area of research, open science is indispensable. Essentially, all of Nong’s simulations rely on free and open-source software (FOSS) at some stage, and many applications of AI or machine learning (ML) utilize open datasets. Nong is a strong proponent of Open Science and Open Education and generally makes software, data, and tutorials generated through her research publicly available under the FAIR (findable, accessible, interoperable, and reusable) principles.

Open Science

One example of Nong’s commitment to open science is the software package ænet, which she released in 2016. It was the first FOSS implementation of machine-learning interatomic potentials (ML-IP). The software is used by researchers worldwide and is available on GitHub ( https://github.com/atomisticnet/aenet ). ænet is well-documented and uses open data formats, such as the plain-text XCrySDen Structure Format (XSF). Its user forum (https://groups.google.com/g/aenet) has many active users, and a Slack workspace is also available for direct discussions and support. In addition to ænet, Nong has published many more GitHub repositories with FOSS for research applications, accessible at https://github.com/atomisticnet .

Nong has made it a general rule to publish data alongside manuscripts. So far, she has published five substantial datasets on the Materials Cloud ( https://www.materialscloud.org ), a platform for Open Science and FAIR sharing of research data. All of Nong’s scientific articles as an independent principal investigator have either been published in open-access journals or are available as preprints on arXiv.

Open Education

In the interest of Open Education, Nong has published approximately ten lectures and tutorials as videos on the YouTube platform. For example, she contributed to the open lecture series "Machine Learning for Materials Science" (ML4MS) in 2019 and the ML-IP workshops in 2021 and 2023, which are fully available online at https://www.mlip-workshop.xyz/speakers . For most tutorials, Nong has also made additional interactive Jupyter notebooks available on GitHub. One example is the machine-learning potential (MLP) “beginner’s guide” that can be downloaded at https://github.com/atomisticnet/MLP-beginners-guide . In 2021, Nong co-organized the virtual Thailand Machine Learning for Chemistry Competition (TMLCC), which spanned three months (September to November) and initially attracted more than 1,000 participants ( https://tmlcc2021.devpost.com ).

To further promote Open Education and Open Infrastructures, Nong represents The Netherlands as a member of the managing committee of the European Union COST Action CA22154 (2024–2028): https://www.cost.eu/actions/CA22154 . This initiative, titled "Data-driven Applications towards the Engineering of Functional Materials: an Open Network" (DAEMON), aims to grow a cross-disciplinary and pan-European network to accelerate materials discovery by providing an open training infrastructure for data science methodologies. Since joining Utrecht University in 2021, Nong has also hosted visiting students and scholars from Japan, Spain, England, Scotland, Belgium, the Basque Country, and Thailand.

June 2021:  
Our comment on "Best Practices in Machine Learning for Chemistry" now appeared online in Nature Chemisty! 
N. Artrith*, K.T. Butler, F.-X. Coudert, S. Han, O. Isayev, A. Jain, A. Walsh, Nat. Chem. 13 (2021) 505-508. Open Access . https://doi.org/10.1038/s41557-021-00716-z