Rethinking Spectral Conditioning for NMR Structure Elucidation

Diffusion models for nuclear magnetic resonance (NMR) structure elucidation typically condition generation through a single global spectral representation, applying the same modulation to every atom and, in graph models, every bond. We study whether this uniform conditioning interface limits structure recovery and replace it with per-element spectral cross-attention, allowing each atom and candidate bond to query the spectral tokens independently. On simulated SpectraBase spectra, this raises top-1 exact recovery on NMR-DiGress from $43.26\%$ for our same-protocol FiLM reproduction to $52.10\%$. The same conditioning change transfers to the coordinate-based ChefNMR architecture, where per-atom cross-attention improves over the released ChefNMR-S model by 7.24 percentage points under a common evaluation. Fine-tuning the cross-attention models with an auxiliary spectral reconstruction loss further reaches $54.38\%$ on NMR-DiGress and $78.29\%$ on ChefNMR. These results show that per-element spectral conditioning improves structure recovery across both graph and coordinate diffusion models.