Metraipolis › archive › dc_05ad401c2fdb

Separability Is Not Alignability: Representations for Spectrum-to-Molecule Retrieval — arXiv:2610.37979

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.37979
License (as archived)
CC0-1.0
Archived
2026-11-01T02:23:19.809Z
SHA-256 of the archived text
283c1cd441087389ee640df161d60c85a69932f665fe50e7c7cd357ed954f923
Stable URL of this page
https://metraipolis.com/documento/dc_05ad401c2fdb

How to cite

Separability Is Not Alignability: Representations for Spectrum-to-Molecule Retrieval — arXiv:2610.37979 — arXiv (en), https://arxiv.org/abs/2610.37979 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_05ad401c2fdb (Metraipolis archive dc_05ad401c2fdb, sha256 283c1cd441087389ee640df161d60c85a69932f665fe50e7c7cd357ed954f923)
BibTeX
@misc{dc05ad401c2fdb,
  title = {Separability Is Not Alignability: Representations for Spectrum-to-Molecule Retrieval — arXiv:2610.37979},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.37979},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_05ad401c2fdb; sha256 283c1cd441087389ee640df161d60c85a69932f665fe50e7c7cd357ed954f923},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.37979. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Identifying a small molecule from its tandem mass spectrum is limited by paired data. Foundation models offer a way around this scarcity: pretrain a spectrum encoder on unlabeled spectra and a molecule encoder on a large corpus of structures, then align the two frozen representations, as in vision-language and as the Platonic Representation Hypothesis suggests. We audit this route on MassSpecGym with spectrum encoders retrained free of test-fold contamination. Alignment needs a metric relating distinct molecules, but the objectives that produce spectrum encoders never see molecular structure: we prove that some predictive objectives fix molecular identity without fixing that metric. Moreover, we find that a candidate-based objective rewards a spectrum-independent prior over the benchmark's decoys, which we isolate. What ultimately addresses data scarcity is not architectural: simulated spectra surpass frozen geometry.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2610.37979 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-11-01T02:12:44.092Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).