发表机构
Shanghai Institute of Microsystem and Information Technology, Chinese Academy of Sciences（中国科学院上海微系统与信息技术研究所）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对光谱缺失或不匹配问题，提出将光谱学领域知识融入混合条件训练，结合专家混合融合进行候选结构重排序，显著提升多模态小分子结构鉴定的鲁棒性。
AI 中文摘要
在实际的分子表征中，小分子结构鉴定受益于互补的光谱证据，但缺失、退化或不匹配的光谱对多模态模型构成挑战。在此，我们将光谱学和化学领域的知识融入混合条件训练，用于候选结构重排序，并采用可复现的评估协议和专家混合（MoE）融合。该协议包含针对每种光谱模态定制的扰动以及基于化学知识的光谱替换，以覆盖光谱可用性、质量和一致性方面的变化。使用来自多模态光谱数据集（MSSD）的模拟光谱，涵盖质谱（MS）、红外光谱（IR）以及1H和13C核磁共振（NMR），在30个预定义条件下共评估了79,462个测试样本，每个样本最多有128个硬候选结构。一项受控的2×2析因比较（完整输入训练与混合条件训练，以及普通拼接与MoE融合，在匹配的评估条件下）表明，混合条件训练在两种架构中都带来了主要增益。对于MoE，平均倒数排名（MRR）在各条件下等权平均后从0.9203提升至0.9763，相对提升6.08%；排名第一的召回率（R@1）在相同条件下平均后从89.50%提升至96.36%，增加了6.86个百分点，相对提升7.67%。仅红外（IR-only）和仅串联质谱（MS/MS-only）的MRR分别从0.4337提升至0.9307和从0.3711提升至0.8575，达到各自基线值的2.15倍和2.31倍，而完整输入的性能保持高位。这些结果支持将领域知识整合到训练条件设计中以提高鲁棒性，并且在混合条件训练下，MoE带来了进一步增益。
英文摘要
Reliable small-molecule identification often requires complementary evidence from multiple spectroscopic measurements. In practice, however, spectra may be unavailable, degraded by measurement-related variations, or even incorrectly associated with a sample, thereby hindering accurate molecular identification. Herein, we propose a multimodal mixed-condition training strategy that accommodates missing, degraded, and mismatched measurements for small-molecule structure identification. The strategy incorporates chemical and spectroscopic knowledge through predefined missing-input configurations, modality-specific spectral perturbations, and chemically informed spectrum replacements. Models were trained on 635,441 samples comprising mass spectrometry (MS), infrared (IR), and nuclear magnetic resonance (NMR) simulated spectra from the Multimodal Spectroscopic Dataset (MSSD). They were then systematically evaluated on 79,462 held-out samples across 30 views designed to represent variations in spectra. A controlled comparison of complete-input and mixed-condition training under concatenation and mixture-of-experts (MoE) fusion showed that the training strategy was the principal source of improvement. For MoE, mixed-condition training increased the mean reciprocal rank (MRR) by 6.08% (from 0.9203 to 0.9763) and the top-1 molecular identification rate by 7.67% (from 89.50% to 96.36%). Notably, under single-modality inputs, IR MRR increased 2.15-fold (from 0.4337 to 0.9307), while MS MRR increased 2.31-fold (from 0.3711 to 0.8575). With the proposed strategy, complete-input performance remained high, while sample-level mismatch detection also improved. Together, these results highlight the potential of multimodal mixed-condition training for practical molecular identification by explicitly addressing incomplete, degraded, and mismatched measurements encountered in real-world analysis.
Comments17 pages, 7 figures, 2 tables. Supplementary information: 12 pages, 3 figures and 11 tables, provided as an ancillary PDF. Revised content
