{
  "abstract": "Foundation models, or large atomistic models (LAMs), aim to universally represent the ground-state potential energy surface (PES) of atomistic systems as defined by density functional theory (DFT). The scaling law is pivotal in the development of large models, suggesting that their generalizability in downstream tasks consistently improves with increased model size, expanded training datasets, and larger computational budgets. In this study, we present DPA3, a multi-layer graph neural network founded on line graph series (LiGS), designed explicitly for the era of LAMs. We demonstrate that the generalization error of the DPA3 model adheres to the scaling law. The scalability in the number of model parameters is attained by stacking additional layers within DPA3. Additionally, the model employs a dataset encoding mechanism that decouples the scaling of training data size from the model size within its multi-task training framework. When trained as problem-oriented potential energy models, the DPA3 model exhibits superior accuracy in the majority of benchmark cases, encompassing systems with diverse features, including molecules, bulk materials, surface and cluster catalysis, two-dimensional materials, and battery materials. When trained as a LAM on the OpenLAM-v1 dataset, the DPA-3.1-3M model exhibits state-of-the-art performance in the LAMBench benchmark suite for LAMs, demonstrating lowest overall zero-shot generalization error across 17 downstream tasks from a broad spectrum of research domains. This performance suggests superior accuracy as an out-of-the-box potential model, requiring minimal fine-tuning data for downstream scientific applications.",
  "abstract_shown": true,
  "affiliations": [
    "AI for Science Institute, Beijing, PR China",
    "Academy for Advanced Interdisciplinary Studies, Peking University, Beijing, PR China",
    "DP Technology, Beijing, PR China",
    "Department of Chemical Engineering, Tsinghua University, Beijing, PR China",
    "State Key Lab of Processors, Institute of Computing Technology, Chinese Academy of Sciences, Beijing, PR China",
    "University of Chinese Academy of Sciences, Beijing, PR China",
    "School of Artificial Intelligence and Data Science, University of Science and Technology of China, Hefei, PR China",
    "Suzhou Big Data & AI Research and Engineering Center, Suzhou, PR China",
    "Suzhou Institute for Advanced Research, University of Science and Technology of China, Suzhou, PR China",
    "School of Chemistry, Sun Yat-sen University, Guangzhou, PR China",
    "Shanghai Engineering Research Center of Molecular Therapeutics & New Drug Development, School of Chemistry and Molecular Engineering, East China Normal University, Shanghai, PR China",
    "College of Chemistry and Molecular Engineering, Peking University, Beijing, PR China",
    "Institute for Advanced algorithms research, Shanghai, PR China",
    "NYU-ECNU Center for Computational Chemistry at NYU Shanghai, Shanghai, PR China",
    "HEDPS, CAPT, College of Engineering, Peking University, Beijing, PR China",
    "National Key Laboratory of Computational Physics, Institute of Applied Physics and Computational Mathematics, Beijing, PR China"
  ],
  "article_type": null,
  "arxiv": null,
  "author_total": 14,
  "authors": [
    {
      "affs": [
        0,
        1,
        2
      ],
      "name": "Duo Zhang",
      "orcid": "0000-0001-9591-2659"
    },
    {
      "affs": [
        0
      ],
      "name": "Anyang Peng",
      "orcid": "0000-0002-0630-2187"
    },
    {
      "affs": [
        0,
        2
      ],
      "name": "Chun Cai",
      "orcid": "0000-0001-6242-0439"
    },
    {
      "affs": [
        3
      ],
      "name": "Wentao Li",
      "orcid": "0000-0002-4522-2163"
    },
    {
      "affs": [
        4,
        5
      ],
      "name": "Yuanchang Zhou"
    },
    {
      "affs": [
        6,
        7,
        8
      ],
      "name": "Jinzhe Zeng",
      "orcid": "0000-0002-1515-8172"
    },
    {
      "affs": [
        0,
        2,
        9
      ],
      "name": "Mingyu Guo",
      "orcid": "0009-0003-2309-0387"
    },
    {
      "affs": [
        0,
        1,
        2
      ],
      "name": "Chengqian Zhang",
      "orcid": "0000-0001-7829-0587"
    },
    {
      "affs": [
        10
      ],
      "name": "Bowen Li",
      "orcid": "0000-0003-4893-0525"
    },
    {
      "affs": [
        11
      ],
      "name": "Hong Yi Jiang",
      "orcid": "0000-0003-3187-2023"
    },
    {
      "affs": [
        12,
        13,
        10
      ],
      "name": "Tong Zhu",
      "orcid": "0000-0001-7472-3736"
    },
    {
      "affs": [
        4,
        5
      ],
      "name": "Weile Jia",
      "orcid": "0000-0001-8539-8326"
    },
    {
      "affs": [
        0,
        2
      ],
      "name": "Linfeng Zhang",
      "orcid": "0000-0002-8470-5846"
    },
    {
      "affs": [
        14,
        15
      ],
      "name": "Han Wang",
      "orcid": "0000-0001-5623-1148"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/0hmf5qp3c7cae.bib",
  "cited_by": 56,
  "doi": "10.1038/s41524-026-02146-2",
  "fulltext_url": null,
  "id": "0hmf5qp3c7cae",
  "issns": [
    "2057-3960"
  ],
  "landing_url": "https://doi.org/10.1038/s41524-026-02146-2",
  "language": "en",
  "licence": "CC-BY-NC-ND",
  "licence_label": "CC BY-NC-ND",
  "licence_source": "journal",
  "licence_url": "https://creativecommons.org/licenses/by-nc-nd/4.0/",
  "metadata_licence": null,
  "native_subjects": [
    [
      "openalex",
      "2505|Machine Learning in Materials Science"
    ]
  ],
  "openalex": "W4417114399",
  "pdf_url": "https://www.nature.com/articles/s41524-026-02146-2_reference.pdf",
  "pmcid": null,
  "pmid": null,
  "published": "2026-05-25",
  "publisher": "Nature Portfolio",
  "ris": "https://research.draconitic.com/paper/0hmf5qp3c7cae.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T13:27:00Z",
      "native_id": "W4417114399",
      "source": "openalex",
      "url": "https://doi.org/10.1038/s41524-026-02146-2"
    }
  ],
  "subjects": [
    {
      "field": "chemistry",
      "origin": "openalex:2505|Machine Learning in Materials Science",
      "subfield": "materials-polymers"
    }
  ],
  "title": "A graph neural network for the era of large atomistic models",
  "title_text": "A graph neural network for the era of large atomistic models",
  "type": "peer-reviewed",
  "updated_at": "2026-11-06T13:20:44Z",
  "url": "https://research.draconitic.com/paper/0hmf5qp3c7cae",
  "venue": "npj Computational Materials",
  "versions": []
}