[22p-E101-1~5] Frontier of applied informatics in the research field of applied physics

[22p-E101-4] Materials discovery using universal Neural Network Potential “Matlantis”

〇Yusuke Asano¹, Hiroki Iriguchi¹, Taku Watanabe², Chikashi Shinagawa³, So Takamoto³ (1.PFCC, 2.ENEOS, 3.PFN)

Keywords:Neural Network Potential, Batteries, Catalysts

In the recent decade, the first-principles calculations have become increasingly important in the design of materials for catalysts, batteries, semiconductors, etc. The neural network potential (NNP) is a technology for predicting the first-principles calculation results at a very fast speed, but it has been used only in narrow target material domains. We have developed an universal NNP for 55 elements, and are providing "Matlantis" as SaaS in order to enable material discovery using NNP in various material fields. In this presentation, we will introduce the technologies of our NNP and examples of material design on "Matlantis".