Bias free multiobjective active learning for materials design and discovery

JSON Export

{
  "revision": 9, 
  "id": "614", 
  "created": "2020-10-26T11:13:42.060539+00:00", 
  "metadata": {
    "doi": "10.24435/materialscloud:8m-6d", 
    "status": "published", 
    "title": "Bias free multiobjective active learning for materials design and discovery", 
    "mcid": "2021.34", 
    "license_addendum": null, 
    "_files": [
      {
        "description": "Features and labels for machine learning (zipped folder of csv files)", 
        "key": "ml_data.zip", 
        "size": 440851, 
        "checksum": "md5:9df8b9ec233d1f551e1e35e21399cf27"
      }, 
      {
        "description": "LAMMPS input files for the calculation of the radii of gyration.", 
        "key": "rg_runs.zip", 
        "size": 436808988, 
        "checksum": "md5:41825511b8ef9210b91814f3b376f810"
      }, 
      {
        "description": "Detailed description of the filecontents.", 
        "key": "README.txt", 
        "size": 1152, 
        "checksum": "md5:4d7e5be16e378137f0aab6f62194f620"
      }, 
      {
        "description": "LAMMPS and SSAGES input files for the calculation of the dimer free energy.", 
        "key": "dimer_runs2.zip", 
        "size": 3680882303, 
        "checksum": "md5:2b7b37bbb2623a6f94469b2a0e20a88d"
      }, 
      {
        "description": "LAMMPS and SSAGES input files for the calculation of the free energy of adsorption.", 
        "key": "adsorption_runs.zip", 
        "size": 3702789205, 
        "checksum": "md5:b8f36997b0b3fbf0898110c30c927d75"
      }
    ], 
    "owner": 70, 
    "_oai": {
      "id": "oai:materialscloud.org:614"
    }, 
    "keywords": [
      "MARVEL", 
      "ERC", 
      "SNSF", 
      "machine learning", 
      "polymers", 
      "multiobjective", 
      "active learning"
    ], 
    "conceptrecid": "613", 
    "is_last": true, 
    "references": [
      {
        "type": "Software", 
        "url": "https://github.com/byooooo/dispersant_screening_PAL", 
        "comment": "Script that can be used to reproduce the main results.", 
        "citation": "K. M. Jablonka, M. J. Giriprasad, S. Wang, B. Smit, and B. Yoo, dispersant_screening_PAL (2020)."
      }, 
      {
        "type": "Software", 
        "url": "https://github.com/kjappelbaum/pypal", 
        "comment": "General-purpose implementation of the active learning algorithm.", 
        "citation": "K. M. Jablonka, M. J. Giriprasad, S. Wang, B. Smit, and B. Yoo, PyPAL (2020)."
      }, 
      {
        "type": "Preprint", 
        "doi": "10.26434/chemrxiv.13200197.v1", 
        "url": "https://chemrxiv.org/articles/preprint/Bias_Free_Multiobjective_Active_Learning_for_Materials_Design_and_Discovery/13200197", 
        "comment": "Preprint where the data is discussed.", 
        "citation": "K. M. Jablonka, M. J. Giriprasad, S. Wang, B. Smit, and B. Yoo, Chemrxiv (2020)."
      }, 
      {
        "type": "Journal reference", 
        "doi": "10.1038/s41467-021-22437-0", 
        "url": "https://www.nature.com/articles/s41467-021-22437-0", 
        "citation": "K. M. Jablonka, G. M. Jothiappan, S. Wang, B. Smit, B. Yoo, Nature Communications 12, 1-10 (2021)"
      }
    ], 
    "publication_date": "Feb 22, 2021, 12:58:24", 
    "license": "Creative Commons Attribution 4.0 International", 
    "id": "614", 
    "description": "The design rules for materials are clear for applications with a single objective. For most applications, however, there are often multiple, sometimes competing objectives where there is no single best material, and the design rules change to finding the set of Pareto optimal materials. \nIn this work, we introduce an active learning algorithm that directly uses the Pareto dominance relation to compute the set of Pareto optimal materials with desirable accuracy. \nWe apply our algorithm to de novo polymer design with a prohibitively large search space.\nUsing molecular simulations, we compute key descriptors for dispersant applications and reduce the number of materials that need to be evaluated to reconstruct the Pareto front with a desired confidence by over 98% compared to random search.\nThis work showcases how simulation and machine learning techniques can be coupled to discover materials within a design space that would be intractable using conventional screening approaches.", 
    "version": 1, 
    "contributors": [
      {
        "email": "kevin.jablonka@epfl.ch", 
        "affiliations": [
          "Laboratory of Molecular Simulation, Institut des Sciences et Ing\u00e9nierie Chimiques, \u00c9cole Polytechnique F\u00e9d\u00e9rale de Lausanne (EPFL), CH-1951 Sion, Valais, Switzerland"
        ], 
        "familyname": "Jablonka", 
        "givennames": "Kevin Maik"
      }, 
      {
        "affiliations": [
          "BASF Corporation, 540 White Plains Road, Tarrytown, New York, 10591, USA"
        ], 
        "familyname": "Melpatti Jothiappan", 
        "givennames": "Giriprasad"
      }, 
      {
        "affiliations": [
          "BASF Corporation, 540 White Plains Road, Tarrytown, New York, 10591, USA"
        ], 
        "familyname": "Wang", 
        "givennames": "Shefang"
      }, 
      {
        "email": "berend.smit@epfl.ch", 
        "affiliations": [
          "Laboratory of Molecular Simulation, Institut des Sciences et Ing\u00e9nierie Chimiques, \u00c9cole Polytechnique F\u00e9d\u00e9rale de Lausanne (EPFL), CH-1951 Sion, Valais, Switzerland"
        ], 
        "familyname": "Smit", 
        "givennames": "Berend"
      }, 
      {
        "email": "brian.yoo@basf.com", 
        "affiliations": [
          "BASF Corporation, 540 White Plains Road, Tarrytown, New York, 10591, USA"
        ], 
        "familyname": "Yoo", 
        "givennames": "Brian"
      }
    ], 
    "edited_by": 70
  }, 
  "updated": "2022-01-23T13:44:03.586692+00:00"
}