{
  "abstract": "Atomic reconstruction reshapes moiré materials across multiple scales, from local structure and polarization textures to global electronic topology, yet direct \\textit{ab initio} modeling becomes prohibitive for large superstructures such as marginal-twist-angle moirés and moiré-of-moirés. We develop MoiréMLIP by fine-tuning a universal atomistic model on the density functional theory labeled Moiré Kaleidoscope dataset, which spans transition metal dichalcogenide compositions, symmetries, stackings, and twist angles. MoiréMLIP reproduces \\textit{ab initio} reconstruction with force errors of 6--8\\,meV/Å and transfers to smaller twist angles and unseen structures. Knowledge distillation yields MoiréMLIP-mini, which retains this accuracy while extending single-GPU inference to one million atoms. Applied to an alternate-twist MoTe$_2$ trilayer, it reveals a hierarchical polarization network spanning tens of nanometers arising from large moiré-cell distortions sharply localized along the moiré-of-moiré domain walls. These results overcome key accuracy, transferability, and scaling bottlenecks of existing atomistic models and enable predictive simulations across emergent moiré length scales.",
  "abstract_shown": true,
  "affiliations": [],
  "article_type": null,
  "arxiv": "2610.04115",
  "author_total": 10,
  "authors": [
    {
      "name": "Thomas Huang"
    },
    {
      "name": "Yueyao Fan"
    },
    {
      "name": "Kaichen Xie"
    },
    {
      "name": "Bolun Li"
    },
    {
      "name": "Kaijie Yang"
    },
    {
      "name": "Jenna A. Bilbrey"
    },
    {
      "name": "Eric Bylaska"
    },
    {
      "name": "Peter V. Sushko"
    },
    {
      "name": "Di Xiao"
    },
    {
      "name": "Ting Cao"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/cd2g1yshwqzbh.bib",
  "cited_by": 0,
  "doi": "10.48550/arxiv.2610.04115",
  "fulltext_url": null,
  "id": "cd2g1yshwqzbh",
  "issns": [],
  "landing_url": "https://arxiv.org/abs/2610.04115",
  "language": null,
  "licence": "LicenseRef-arXiv-nonexclusive-1.0",
  "licence_label": "arXiv licence",
  "licence_source": "article",
  "licence_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
  "metadata_licence": "CC0-1.0",
  "native_subjects": [
    [
      "arxiv",
      "cond-mat.mes-hall"
    ],
    [
      "arxiv",
      "cond-mat.mtrl-sci"
    ],
    [
      "openalex",
      "2505|Machine Learning in Materials Science"
    ]
  ],
  "openalex": "W7220714069",
  "pdf_url": "https://arxiv.org/pdf/2610.04115",
  "pmcid": null,
  "pmid": null,
  "published": "2026-10-02",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/cd2g1yshwqzbh.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:27:01Z",
      "native_id": "2610.04115",
      "source": "arxiv",
      "url": "https://arxiv.org/abs/2610.04115"
    }
  ],
  "subjects": [
    {
      "field": "physics",
      "origin": "arxiv:cond-mat.mes-hall",
      "subfield": "condensed-matter"
    }
  ],
  "title": "Distilling universal machine-learning potentials for moiré lattices across one million atoms",
  "title_text": "Distilling universal machine-learning potentials for moiré lattices across one million atoms",
  "type": "preprint",
  "updated_at": "2026-10-30T12:27:01Z",
  "url": "https://research.draconitic.com/paper/cd2g1yshwqzbh",
  "venue": "arXiv",
  "versions": []
}