Quanvolutional Neural Networks for Spectrum Peak-Finding

2025-12-15

AI Summary

Peak identification and localization (deconvolution) in complex spectra—such as nuclear magnetic resonance (NMR)—remains a challenging yet critical task in analytical chemistry. Method: This paper introduces QuanvNN, the first interpretable and lightweight quantum convolutional neural network specifically designed for spectral peak detection. It employs parameterized quantum circuits to implement quanvolutional layers within a classical–quantum hybrid architecture and is trained end-to-end on a synthetic NMR dataset for joint tasks of peak counting and position regression. Contribution/Results: Compared to an equivalent classical CNN, QuanvNN achieves significantly improved convergence stability on hard samples and superior generalization: +11% in F1-score and −30% in mean absolute error for peak position estimation. This work provides the first empirical validation of quantum convolution’s efficacy and practicality in spectral analysis, establishing a novel paradigm for deploying quantum machine learning in chemometrics.

Technology Category

• Machine Learning: Quantum Machine Learning
• Search and Optimization: Non-convex Optimization
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• Search and Retrieval-Augmented AI: Web query analysis, representation and understanding
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work

Abstract

The analysis of spectra, such as Nuclear Magnetic Resonance (NMR) spectra, for the comprehensive characterization of peaks is a challenging task for both experts and machines, especially with complex molecules. This process, also known as deconvolution, involves identifying and quantifying the peaks in the spectrum. Machine learning techniques have shown promising results in automating this process. With the advent of quantum computing, there is potential to further enhance these techniques. In this work, inspired by the success of classical Convolutional Neural Networks (CNNs), we explore the use of Quanvolutional Neural Networks (QuanvNNs) for the multi-task peak finding problem, involving both peak counting and position estimation. We implement a simple and interpretable QuanvNN architecture that can be directly compared to its classical CNN counterpart, and evaluate its performance on a synthetic NMR-inspired dataset. Our results demonstrate that QuanvNNs outperform classical CNNs on challenging spectra, achieving an 11% improvement in F1 score and a 30% reduction in mean absolute error for peak position estimation. Additionally, QuanvNNs appear to exhibit better convergence stability for harder problems.

Problem

• Automating NMR spectrum peak identification and quantification
• Enhancing peak counting and position estimation with quantum computing
• Improving accuracy and stability in spectral deconvolution tasks

Innovation

• Quanvolutional Neural Networks for peak finding
• Interpretable architecture compared to classical CNNs
• Outperforms CNNs in F1 score and error reduction