Conditional Generative Framework with Peak-Aware Attention for Robust Chemical Detection under Interferences

TL;DR

This work tackles robust chemical detection with GC-MS under interference by introducing a peak-aware attention mechanism and a conditional GAN that encodes solvent and target chemical information into latent representations. The CGAN generates faithful synthetic GC-MS spectra that preserve diagnostic peaks, while a peak-aware discriminator-guided detector leverages both real and synthetic data to improve discrimination accuracy. Quantitative results show global spectral fidelity (Cosine similarity and PCC > 0.94) and strong improvements in detection performance as synthetic data volume increases, converging near optimal levels beyond 615 samples. The approach reduces experimental data requirements and enhances practical robustness of chemical identification in complex, real-world interference scenarios.

Abstract

Gas chromatography-mass spectrometry (GC-MS) is a widely used analytical method for chemical substance detection, but measurement reliability tends to deteriorate in the presence of interfering substances. In particular, interfering substances cause nonspecific peaks, residence time shifts, and increased background noise, resulting in reduced sensitivity and false alarms. To overcome these challenges, in this paper, we propose an artificial intelligence discrimination framework based on a peak-aware conditional generative model to improve the reliability of GC-MS measurements under interference conditions. The framework is learned with a novel peak-aware mechanism that highlights the characteristic peaks of GC-MS data, allowing it to generate important spectral features more faithfully. In addition, chemical and solvent information is encoded in a latent vector embedded with it, allowing a conditional generative adversarial neural network (CGAN) to generate a synthetic GC-MS signal consistent with the experimental conditions. This generates an experimental dataset that assumes indirect substance situations in chemical substance data, where acquisition is limited without conducting real experiments. These data are used for the learning of AI-based GC-MS discrimination models to help in accurate chemical substance discrimination. We conduct various quantitative and qualitative evaluations of the generated simulated data to verify the validity of the proposed framework. We also verify how the generative model improves the performance of the AI discrimination framework. Representatively, the proposed method is shown to consistently achieve cosine similarity and Pearson correlation coefficient values above 0.9 while preserving peak number diversity and reducing false alarms in the discrimination model.

Figures (5)

• Figure 1: Comparison of GC-MS spectral representations. (a) Original experimental GC-MS data of 2-CEES with ethanol solvent. (b)--(d) Limitations of artificial intelligence-based generative models based on GC-MS data characteristics (TimeGAN yoon2019time, LSTM-CNN GAN zhu2019electrocardiogram, DCGAN dewi2022synthetic).
• Figure 2: GC-MS Detection Model Framework Using peak-aware attention-Based Conditional Generation Model.
• Figure 3: GC data of chemicals using four solvents.
• Figure 4: Representative visual comparison of GC-MS spectra between real and generated data. (a,b) shows that it preserves the peak position and intensity pattern through the case of 4-nitrophenol + MC. (c,d) shows that the proposed peak-aware attention based generative model successfully reconstructs fine-grained peaks even with a complex multi-agent scenario,2-CEES + 2-CEPS + DFP + DMMP + MC.
• Figure 5: Changes in detection performance with synthetic training data volume.