E(3)-equivariant graph neural networks for data-efficient and accurate interatomic potentials

JSON Export

{
  "metadata": {
    "edited_by": 576, 
    "owner": 720, 
    "description": "This work presents Neural Equivariant Interatomic Potentials (NequIP), an E(3)-equivariant neural network approach for learning interatomic potentials from ab-initio calculations for molecular dynamics simulations. While most contemporary symmetry-aware models use invariant convolutions and only act on scalars, NequIP employs E(3)-equivariant convolutions for interactions of geometric tensors, resulting in a more information-rich and faithful representation of atomic environments. The method achieves state-of-the-art accuracy on a challenging and diverse set of molecules and materials while exhibiting remarkable data efficiency. NequIP outperforms existing models with up to three orders of magnitude fewer training data, challenging the widely held belief that deep neural networks require massive training sets. The high data efficiency of the method allows for the construction of accurate potentials using high-order quantum chemical level of theory as reference and enables high-fidelity molecular dynamics simulations over long time scales.", 
    "keywords": [
      "machine learning", 
      "molecular dynamics", 
      "density-functional theory"
    ], 
    "is_last": true, 
    "title": "E(3)-equivariant graph neural networks for data-efficient and accurate interatomic potentials", 
    "status": "published", 
    "license_addendum": null, 
    "doi": "10.24435/materialscloud:s0-5n", 
    "conceptrecid": "1301", 
    "_files": [
      {
        "checksum": "md5:d436cfb166276304c8acf9ffe8f36738", 
        "description": "Li4P2O7 quench data set", 
        "key": "lipo-quench.xyz", 
        "size": 550397190
      }, 
      {
        "checksum": "md5:16f91b4aca0b5759026f8bd62790e5c6", 
        "description": "Li6.75P3S11 data set", 
        "key": "lips.xyz", 
        "size": 222930996
      }, 
      {
        "checksum": "md5:eb0224ff48b6cf3592ec96de0a1548b2", 
        "description": "Formate on Cu dataset", 
        "key": "fcu.xyz", 
        "size": 38354321
      }, 
      {
        "checksum": "md5:aa22b3ba42b42ed12fc05dea78b6ba01", 
        "description": "README file.", 
        "key": "README.md", 
        "size": 558
      }
    ], 
    "references": [
      {
        "citation": "Batzner, S., Musaelian, A., Sun, L., Geiger, M., Mailoa, J. P., Kornbluth, M., ... & Kozinsky, B. (2021). E(3)-equivariant graph neural networks for data-efficient and accurate interatomic potentials. arXiv preprint arXiv:2101.03164.", 
        "comment": "Preprint where the data is discussed", 
        "url": "https://arxiv.org/abs/2101.03164", 
        "doi": "", 
        "type": "Preprint"
      }
    ], 
    "contributors": [
      {
        "givennames": "Simon", 
        "affiliations": [
          "John A. Paulson School of Engineering and Applied Sciences, Harvard University, Cambridge, MA 02138, USA"
        ], 
        "familyname": "Batzner", 
        "email": "batzner@g.harvard.edu"
      }, 
      {
        "givennames": "Albert", 
        "affiliations": [
          "John A. Paulson School of Engineering and Applied Sciences, Harvard University, Cambridge, MA 02138, USA"
        ], 
        "familyname": "Musaelian", 
        "email": "albym@seas.harvard.edu"
      }, 
      {
        "givennames": "Lixin", 
        "affiliations": [
          "John A. Paulson School of Engineering and Applied Sciences, Harvard University, Cambridge, MA 02138, USA"
        ], 
        "familyname": "Sun", 
        "email": "lixinsun@g.harvard.edu"
      }, 
      {
        "givennames": "Mario", 
        "affiliations": [
          "Research Laboratory of Electronics, Massachusetts Institute of Technology, Cambridge, MA 02139, USA"
        ], 
        "familyname": "Geiger", 
        "email": "geiger.mario@gmail.com"
      }, 
      {
        "givennames": "Jonathan P.", 
        "affiliations": [
          "Robert Bosch Research and Technology Center, Cambridge, MA 02139, USA"
        ], 
        "familyname": "Mailoa", 
        "email": "jpmailoa@gmail.com"
      }, 
      {
        "givennames": "Mordechai", 
        "affiliations": [
          "Robert Bosch Research and Technology Center, Cambridge, MA 02139, USA"
        ], 
        "familyname": "Kornbluth", 
        "email": "Mordechai.Kornbluth@us.bosch.com"
      }, 
      {
        "givennames": "Nicola", 
        "affiliations": [
          "John A. Paulson School of Engineering and Applied Sciences, Harvard University, Cambridge, MA 02138, USA"
        ], 
        "familyname": "Molinari", 
        "email": "nmolinari@seas.harvard.edu"
      }, 
      {
        "givennames": "Tess E.", 
        "affiliations": [
          "Computational Research Division and Center for Advanced Mathematics for Energy Research Applications, Lawrence Berkeley National Laboratory, Berkeley, CA 94720, USA", 
          "Department of Electrical Engineering and Computer Science and Research Laboratory of Electronics, Massachusetts Institute of Technology, Cambridge, MA 02139, USA"
        ], 
        "familyname": "Smidt", 
        "email": "tsmidt@mit.edu"
      }, 
      {
        "givennames": "Boris", 
        "affiliations": [
          "John A. Paulson School of Engineering and Applied Sciences, Harvard University, Cambridge, MA 02138, USA", 
          "Robert Bosch Research and Technology Center, Cambridge, MA 02139, USA"
        ], 
        "familyname": "Kozinsky", 
        "email": "bkoz@seas.harvard.edu"
      }
    ], 
    "_oai": {
      "id": "oai:materialscloud.org:1302"
    }, 
    "publication_date": "Mar 30, 2022, 16:37:02", 
    "mcid": "2022.45", 
    "version": 1, 
    "id": "1302", 
    "license": "Creative Commons Attribution 4.0 International"
  }, 
  "revision": 5, 
  "created": "2022-03-29T20:55:42.981128+00:00", 
  "id": "1302", 
  "updated": "2022-03-30T14:37:02.527078+00:00"
}