{
  "abstract": "Molecular line lists disagree at a level that measurably shifts retrieved C/O, so retrievals conditioned on a single list omit a dominant systematic from their error bars. Classical marginalization re-runs a full retrieval under K opacity realizations per object. We move that cost into simulation: while generating the training set we randomize opacities (smooth multiplicative Fourier fields in $\\logν$ plus a discrete line-list mixture, under a deliberately broad prior informed by measured inter-list differences), so a neural posterior estimator learns an already-marginalized posterior. Conceptually, this is the amortized analogue of repeating a conventional retrieval under many plausible opacity realizations and averaging the resulting posteriors. On perturbed held-out spectra (n=2048) a standard NPE covers nominal 90% C/O intervals 43.3% of the time against 82.4% for a compute-matched marginalized network; three-seed control ensembles reach 93.2% on clean data but 61.4% under perturbation, against 89.4% marginalized, isolating the deficit as opacity-induced (TARP joint expected-coverage deviation 0.005 against 0.035 for a single control network). An axis ablation traces the gain to the continuous field (45.4% for the discrete swaps alone, 79.3% for the field alone). Applying the 105 real bad-pixel masks with median filling then collapsed the ensemble's C/O coverage from 0.901 to 0.479; retraining on the same mask distribution and pooling three such seeds returns a nominal 0.905 C/O and 0.891 overall under the full deployment path, still in simulation (n=1024, same 105 masks), at 0.21 to 0.31 of the prior width. On 105 JWST/G395H spectral products of 13 late-T/Y dwarfs, single marginalized networks bracket the Hood et al. (2024) T8 benchmark C/O where the control excludes it, a consistency check rather than a validation, at 1.6-2.6x wider intervals.",
  "abstract_shown": true,
  "affiliations": [],
  "article_type": null,
  "arxiv": "2609.01665",
  "author_total": 1,
  "authors": [
    {
      "name": "Kellen Heraty"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/4h6becfk916ys.bib",
  "cited_by": 0,
  "doi": "10.48550/arxiv.2609.01665",
  "fulltext_url": null,
  "id": "4h6becfk916ys",
  "issns": [],
  "landing_url": "https://arxiv.org/abs/2609.01665",
  "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",
      "astro-ph.EP"
    ],
    [
      "arxiv",
      "astro-ph.IM"
    ],
    [
      "arxiv",
      "astro-ph.SR"
    ],
    [
      "openalex",
      "3103|Stellar, planetary, and galactic studies"
    ]
  ],
  "openalex": "W7207675914",
  "pdf_url": "https://arxiv.org/pdf/2609.01665",
  "pmcid": null,
  "pmid": null,
  "published": "2026-10-05",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/4h6becfk916ys.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:27:01Z",
      "native_id": "2609.01665",
      "source": "arxiv",
      "url": "https://arxiv.org/abs/2609.01665"
    }
  ],
  "subjects": [
    {
      "field": "astronomy",
      "origin": "arxiv:astro-ph.IM",
      "subfield": "instrumentation-methods"
    },
    {
      "field": "astronomy",
      "origin": "arxiv:astro-ph.EP",
      "subfield": "planetary"
    },
    {
      "field": "astronomy",
      "origin": "arxiv:astro-ph.SR",
      "subfield": "stellar"
    }
  ],
  "title": "Amortized Opacity Marginalization Improves C/O Interval Calibration for Brown-Dwarf Retrievals",
  "title_text": "Amortized Opacity Marginalization Improves C/O Interval Calibration for Brown-Dwarf Retrievals",
  "type": "preprint",
  "updated_at": "2026-10-30T12:27:01Z",
  "url": "https://research.draconitic.com/paper/4h6becfk916ys",
  "venue": "arXiv",
  "versions": []
}