{
  "abstract": "Most deep learning models for 1H NMR are trained for a single endpoint and do not provide representations that can be reused across analytical tasks. We introduce ROSE (Reusable One-dimensional Spectrum Embeddings), a 7.8-million-parameter spectrum-side foundation model pretrained on 3.2 million experimental and simulated spectra using an InChIKey-14 identity holdout that excludes all molecules appearing in the external benchmark test sets. Its shared spectral encoder supports denoising, peak detection, classification, cross-field alignment, molecular retrieval and structure-tospectrum prediction. Beyond internal evaluation on identity-disjoint holdouts, ROSE was assessed on external benchmarks excluded from pretraining, matching the published QIB PLS-DA result for 60-MHz edible-oil classification (98.8% balanced accuracy vs 98.9% accuracy), reducing NMRNet structure-to-spectrum Chamfer loss by 39% over Morgan+Ridge (0.79 vs 1.29), achieving 69% peak-detection F1 on NMRformer, and, after short domain adaptation, reaching 37.1% Top-1 and 67.7% Top-10 retrieval on NMR-Solver. ROSE represents a first step toward a general-purpose NMR foundation model for scalable industrial analytical applications.",
  "abstract_shown": true,
  "affiliations": [
    "Italy",
    "Università di Cagliari"
  ],
  "article_type": null,
  "arxiv": null,
  "author_total": 4,
  "authors": [
    {
      "affs": [
        0
      ],
      "name": "Mattia Di Iorio"
    },
    {
      "affs": [
        0
      ],
      "name": "Carmine Mattia"
    },
    {
      "affs": [
        0
      ],
      "name": "Andrea Zanda",
      "orcid": "0009-0004-2867-9374"
    },
    {
      "affs": [
        1
      ],
      "name": "Maurizio Atzori"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/72x1yzfcw659a.bib",
  "cited_by": 0,
  "doi": "10.26434/chemrxiv.15007823/v2",
  "fulltext_url": null,
  "id": "72x1yzfcw659a",
  "issns": [],
  "landing_url": "https://chemrxiv.org/doi/full/10.26434/chemrxiv.15007823/v2",
  "language": null,
  "licence": "CC-BY",
  "licence_label": "CC BY",
  "licence_source": "journal",
  "licence_url": "https://creativecommons.org/licenses/by/4.0/",
  "metadata_licence": null,
  "native_subjects": [
    [
      "openalex",
      "1602|Spectroscopy and Chemometric Analyses"
    ]
  ],
  "openalex": "W7215085766",
  "pdf_url": "https://chemrxiv.org/doi/pdf/10.26434/chemrxiv.15007823/v2",
  "pmcid": null,
  "pmid": null,
  "published": "2026-09-30",
  "publisher": "American Chemical Society (ACS)",
  "ris": "https://research.draconitic.com/paper/72x1yzfcw659a.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:17:21Z",
      "native_id": "10.26434/chemrxiv.15007823",
      "source": "chemrxiv",
      "url": "https://chemrxiv.org/doi/full/10.26434/chemrxiv.15007823/v2"
    }
  ],
  "subjects": [
    {
      "field": "chemistry",
      "origin": "openalex:1602|Spectroscopy and Chemometric Analyses",
      "subfield": "analytical"
    }
  ],
  "title": "ROSE: a Foundation Model for Reusable One-dimensional Spectrum Embeddings in 1H NMR",
  "title_text": "ROSE: a Foundation Model for Reusable One-dimensional Spectrum Embeddings in 1H NMR",
  "type": "preprint",
  "updated_at": "2026-10-30T12:17:21Z",
  "url": "https://research.draconitic.com/paper/72x1yzfcw659a",
  "venue": "ChemRxiv",
  "versions": []
}