{
  "abstract": "When a graph neural network estimates a physical quantity, its frame, edge direction, aggregation and feature normalization decide which laws it can break, and wrong choices leave systematic errors the loss never registers. We build a linac beam-dynamics surrogate on this observation. Space charge, short-range wakefields and coherent synchrotron radiation enter as bounded closures on analytic kernels; the graphs carry a frame and a direction dictated by the interaction; and their symmetry makes the net space-charge self-force vanish for any weights. Verified without training, with randomized weights, the invariants break by five to eight orders of magnitude once the responsible feature is inverted. Per-working-point models constrained by the rms envelope equation are distilled into one model conditioned on 15 machine settings. On the 18.4~m L0+L1 section of the Shanghai soft x-ray free-electron laser linac, this model predicts the transverse beam envelope at 50 unseen settings to about $7\\%$, together with the centroid trajectory and emittance, in 0.38~s on a single Quadro RTX 6000 GPU. Training is instead dominated by a normalization choice, a data-derived output range for the beam size: at all 16 working points it lowers the starting error 2.2- to 8.5-fold and fixes the convergence speed. The remaining accuracy differences between hybrid and data-driven architectures vary with the working point and with the density of observation. The held-out error scales roughly as $N^{-0.6}$ with the corpus size $N$. Whether the physics pays off where collective effects are strong, as in a magnetic bunch compressor, is in the next phase of our study.",
  "abstract_shown": true,
  "affiliations": [],
  "article_type": null,
  "arxiv": "2610.28250",
  "author_total": 6,
  "authors": [
    {
      "name": "Bowen Zhang"
    },
    {
      "name": "Dazhang Huang"
    },
    {
      "name": "Nengyuan Zhang"
    },
    {
      "name": "Duan Gu"
    },
    {
      "name": "Chao Feng"
    },
    {
      "name": "Bo Liu"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/8gq79xs21vy52.bib",
  "cited_by": 0,
  "doi": "10.48550/arxiv.2610.28250",
  "fulltext_url": null,
  "id": "8gq79xs21vy52",
  "issns": [],
  "landing_url": "https://arxiv.org/abs/2610.28250",
  "language": null,
  "licence": "CC-BY-4.0",
  "licence_label": "CC BY 4.0",
  "licence_source": "article",
  "licence_url": "http://creativecommons.org/licenses/by/4.0/",
  "metadata_licence": "CC0-1.0",
  "native_subjects": [
    [
      "arxiv",
      "physics.acc-ph"
    ],
    [
      "openalex",
      "2202|Particle accelerators and beam dynamics"
    ]
  ],
  "openalex": "W7223116592",
  "pdf_url": "https://arxiv.org/pdf/2610.28250",
  "pmcid": null,
  "pmid": null,
  "published": "2026-10-25",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/8gq79xs21vy52.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:35:35Z",
      "native_id": "2610.28250",
      "source": "arxiv",
      "url": "https://arxiv.org/abs/2610.28250"
    }
  ],
  "subjects": [
    {
      "field": "physics",
      "origin": "arxiv:physics.acc-ph",
      "subfield": "applied-instrumentation"
    }
  ],
  "title": "Physics-constrained graph closures and working-point conditioning for a fast surrogate of linac beam dynamics",
  "title_text": "Physics-constrained graph closures and working-point conditioning for a fast surrogate of linac beam dynamics",
  "type": "preprint",
  "updated_at": "2026-10-30T12:35:35Z",
  "url": "https://research.draconitic.com/paper/8gq79xs21vy52",
  "venue": "arXiv",
  "versions": []
}