Recent progress in the NIST-JARVIS infrastructure

[Kamal Choudhary]

Dear JARVIS users,

 

We wanted to thank you for your continued support of the JARVIS project at NIST. It is our great pleasure to share that the new JARVIS review article has been published in the Applied Physics Reviews: “Recent progress in the JARVIS infrastructure for next-generation data-driven materials design”, Appl. Phys. Rev. 10, 041302 (2023) https://doi.org/10.1063/5.0159299

 

A summary of the recent updates to JARVIS include:

1) doubling the number of materials in the database since its first release (80,000),

2) incorporating more accurate electronic structure methods (Quantum Monte Carlo),

3) including graph neural network-based materials design (ALIGNN),

4) development of unified force-field (ALIGNN-FF),

5) development of a universal tight-binding model (ThreeBodyTB),

6) addition of computer-vision tools for advanced microscopy applications (AtomVision),

7) development of a natural language processing tool for text-generation and analysis (ChemNLP),

8) debuting a large-scale benchmarking endeavor (JARVIS-Leaderboard),

9) including quantum computing algorithms for solids (AtomQC),

10) integrating several experimental datasets,

11) staging several community engagement and outreach events (AIMS and QMMS workshops, JARVIS-Schools)

 

If you use JARVIS for your research, we ask that you please cite the following two articles:

1. npj Computational Materials 6, 173 (2020) https://doi.org/10.1038/s41524-020-00440-1
2. Appl. Phys. Rev. 10, 041302 (2023) https://doi.org/10.1063/5.0159299

 

If you would like to have a JARVIS-School at your institute please fill up this google-form. 

If you like a new feature added in JARVIS, please feel free to raise a GitHub issue.

We thank you again for being a part of the JARVIS community and look forward to growing this project further in the coming years. 

Best regards,

The NIST-JARVIS team