arXiv:2610.37979 · v1
Separability Is Not Alignability: Representations for Spectrum-to-Molecule Retrieval
摘要
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.
作者备注：22 pages, 3 figures. Accepted as a poster at the NeurIPS 2026 Workshop on Representations for the Physical Sciences (RPS). OpenReview: https://openreview.net/forum?id=kqgbNulQ8A