학술논문
A foundation model for atomistic materials chemistry
Document Type
Working Paper
Author
Batatia, Ilyes; Benner, Philipp; Chiang, Yuan; Elena, Alin M.; Kovács, Dávid P.; Riebesell, Janosh; Advincula, Xavier R.; Asta, Mark; Avaylon, Matthew; Baldwin, William J.; Berger, Fabian; Bernstein, Noam; Bhowmik, Arghya; Blau, Samuel M.; Cărare, Vlad; Darby, James P.; De, Sandip; Della Pia, Flaviano; Deringer, Volker L.; Elijošius, Rokas; El-Machachi, Zakariya; Falcioni, Fabio; Fako, Edvin; Ferrari, Andrea C.; Genreith-Schriever, Annalena; George, Janine; Goodall, Rhys E. A.; Grey, Clare P.; Grigorev, Petr; Han, Shuang; Handley, Will; Heenen, Hendrik H.; Hermansson, Kersti; Holm, Christian; Jaafar, Jad; Hofmann, Stephan; Jakob, Konstantin S.; Jung, Hyunwook; Kapil, Venkat; Kaplan, Aaron D.; Karimitari, Nima; Kermode, James R.; Kroupa, Namu; Kullgren, Jolla; Kuner, Matthew C.; Kuryla, Domantas; Liepuoniute, Guoda; Margraf, Johannes T.; Magdău, Ioan-Bogdan; Michaelides, Angelos; Moore, J. Harry; Naik, Aakash A.; Niblett, Samuel P.; Norwood, Sam Walton; O'Neill, Niamh; Ortner, Christoph; Persson, Kristin A.; Reuter, Karsten; Rosen, Andrew S.; Schaaf, Lars L.; Schran, Christoph; Shi, Benjamin X.; Sivonxay, Eric; Stenczel, Tamás K.; Svahn, Viktor; Sutton, Christopher; Swinburne, Thomas D.; Tilly, Jules; van der Oord, Cas; Varga-Umbrich, Eszter; Vegge, Tejs; Vondrák, Martin; Wang, Yangshuai; Witt, William C.; Zills, Fabian; Csányi, Gábor
Source
Subject
Language
Abstract
Machine-learned force fields have transformed the atomistic modelling of materials by enabling simulations of ab initio quality on unprecedented time and length scales. However, they are currently limited by: (i) the significant computational and human effort that must go into development and validation of potentials for each particular system of interest; and (ii) a general lack of transferability from one chemical system to the next. Here, using the state-of-the-art MACE architecture we introduce a single general-purpose ML model, trained on a public database of 150k inorganic crystals, that is capable of running stable molecular dynamics on molecules and materials. We demonstrate the power of the MACE-MP-0 model - and its qualitative and at times quantitative accuracy - on a diverse set problems in the physical sciences, including the properties of solids, liquids, gases, chemical reactions, interfaces and even the dynamics of a small protein. The model can be applied out of the box and as a starting or "foundation model" for any atomistic system of interest and is thus a step towards democratising the revolution of ML force fields by lowering the barriers to entry.
Comment: 119 pages, 63 figures, 37MB PDF
Comment: 119 pages, 63 figures, 37MB PDF