SpecCal：面向红外光谱分子结构重建的歧义感知候选校准方法
SpecCal: Ambiguity-Aware Candidate Calibration for Infrared Spectrum-Based Molecular Structure Reconstruction
- College of Computer Science and Technology, Jilin University（吉林大学计算机科学与技术学院）
- School of Mathematics, Jilin University（吉林大学数学学院）
- Information Hub, HKUST(GZ)（香港科技大学（广州）信息中心）
- School of Chemistry, Chemical Engineering and Biotechnology (CCEB), Nanyang Technological University（南洋理工大学化学、化学工程与生物工程学院）
- School of Software Engineering, University of Science and Technology of China（中国科学技术大学软件学院）
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
本研究针对红外光谱分子结构重建的歧义问题，提出即插即用、与模型无关的SpecCal校准框架，可提升不同基础模型在SMILES和骨架水平的top-k重建性能。
AI中文摘要：
从红外（IR）光谱推断分子结构是一项基础但极具挑战性的任务，核心难点在于红外光谱提供的结构信息有限：不同分子可能共享相似的官能团和局部振动模式，导致光谱响应高度相似。因此，即使观测到的光谱对应唯一的潜在结构，从该光谱重建结构仍存在歧义。现有红外转分子模型通常生成一个排序后的候选分子集，但该集合很大程度上由模型学习到的生成偏好决定，可能未充分捕捉最符合观测光谱约束的结构。为解决这一局限，我们提出SpecCal，一种用于红外转分子预测的无训练候选校准框架。SpecCal基于现有基础模型的候选输出进行操作，通过重新排序当前候选并引入由光谱一致性引导的结构上合理的替代方案来改进预测集。该框架即插即用且与模型无关，与不同基础模型集成无需参数更新。在多个基准上的实验表明，SpecCal在不同基础模型上始终提升SMILES和骨架水平的top-k重建性能。进一步分析显示，在光谱歧义下校准候选集为改进红外光谱的分子重建提供了实用方法，代码可获取于：this https URL。
英文摘要：
Inferring molecular structures from infrared (IR) spectra is a fundamental yet challenging problem. A key difficulty is that an IR spectrum provides limited structural information: different molecules may share similar functional groups and local vibrational patterns, leading to highly similar spectral responses. Thus, even when an observed spectrum has a unique underlying structure, reconstructing it from the spectrum remains ambiguous. Existing IR-to-molecule models usually generate a ranked set of candidate molecules, but this set is largely determined by the model's learned generation preference and may not fully capture the structures that best satisfy the observed spectral constraints. To address this limitation, we propose SpecCal, a training-free candidate calibration framework for IR-to-molecule prediction. SpecCal operates on the candidate outputs of existing base models and improves the prediction set by re-ranking current candidates while introducing additional structurally plausible alternatives guided by spectral consistency. The framework is plug-and-play and model-agnostic, requiring no parameter updates for integration with diverse base models. Experiments on multiple benchmarks show that SpecCal consistently improves top-k reconstruction at both SMILES and scaffold levels across different base models. Further analyses demonstrate that calibrating candidate sets under spectral ambiguity provides a practical way to improve molecular reconstruction from IR spectra. The code is available at: https://anonymous.4open.science/r/SpecCal-B18A.
