Paradigm shift in nuclear magnetic resonance relaxation spectra inversion: The relay and integration of regularization, machine learning, and quantum computing
Abstract
Nuclear magnetic resonance technology has been widely applied in deep geo-energy fields, including deep earth oil and gas exploration, geothermal resource exploration, carbon dioxide geological sequestration, and underground hydrogen storage. The reliability of these applications is directly affected by the accuracy of nuclear magnetic resonance relaxation spectra inversion. However, the inverse problem is inherently ill-posed and therefore the effective inversion strategy with appropriate constraints is required. The evolution of nuclear magnetic resonance relaxation spectra inversion has been marked by three major methodological milestones. The classical regularization method, exemplified by Tikhonov regularization, establishes the mathematical foundation; the machine-learning approach, particularly deep neural networks, is reshaping the computational paradigm; and the quantum computing offers a forward-looking avenue for fundamentally overcoming the computational-complexity bottleneck associated with inversion. This work emphasizes that the three paradigms are not characterized by simple linear replacement, but rather by the succession and integration. Each generation builds on its predecessor and extends the boundaries of nuclear magnetic resonance inversion. Specifically, regularization provides mathematical rigor, machine learning enables data-driven adaptability, and quantum computing may unlock fundamental computational speedups. This perspective systematically describes this evolutionary trajectory and discusses future directions for the integration of these paradigms.
Document Type: Perspective
Cited as: Guo, J., Wan, Q., Li, Q., Xie, R., Xiao, L. Paradigm shift in nuclear magnetic resonance relaxation spectra inversion: The relay and integration of regularization, machine learning, and quantum computing. Advances in Geo-Energy Research, 2026, 22(1): 87-90. https://doi.org/10.46690/ager.2026.10.07
DOI:
https://doi.org/10.46690/ager.2026.10.07Keywords:
Nuclear magnetic resonance, relaxation spectra inversion, regularization, machine learning, quantum computingReferences
Ahola, S., Zhivonitko, V. V., Mankinen, O., et al. Ultrafast multidimensional Laplace NMR for a rapid and sensitive chemical analysis. Nature Communications, 2015, 6(1): 8363.
Arridge, S., Maass, P., Öktem, O., et al. Solving inverse problems using data-driven models. Acta Numerica, 2019, 28: 1-174.
Butler, J. P., Reeds, J. A., Dawson, S. V. Estimating solutions of first kind integral equations with nonnegative constraints and optimal smoothing. SIAM Journal on Numerical Analysis, 1981, 18(3): 381-397.
Chen, B., Zhang, Y., Wang, L., et al. High-confidence reconstruction for Laplace inversion in NMR based on uncertainty-informed deep learning. Science Advances, 2025, 11(35): eadw1379.
Guo, J., Fan, J., Rui, Z., et al. A comprehensive review of crude oil viscosity characterization with nuclear magnetic resonance technique. Journal of Geophysics and Engineering, 2026a, 23(4): 1249-1270.
Guo, J., Wan, Q., Wang, Y., et al. Numerical investigations on nuclear magnetic resonance T₁–T₂ responses in tight gas reservoirs with oil-based mud filtrate invasion. SPE Journal, 2026b, 31(6): 3910-3925.
Guo, J., Zhao, Y., Xie, R., et al. Enhancement of NMR relaxation inversion-A review on pretreatment denoising. Petroleum Science, 2026c, 23(7): 3947-3971.
Guo, J., Xie, R., Zou, Y., et al. A new method for NMR data inversion based on double-parameter regularization. Geophysics, 2018, 83(5): JM39-JM49.
Harrow, A. W., Hassidim, A., Lloyd, S. Quantum algorithm for linear systems of equations. Physical Review Letters, 2009, 103(15): 150502.
Ilmavirta, J., Lassas, M., Lu, J., et al. Quantum computing algorithms for inverse problems on graphs and an NP-complete inverse problem. Inverse Problems and Imaging, 2025, 19(4): 660-692.
Lamanna, R. On the inversion of multicomponent NMR relaxation and diffusion decays in heterogeneous systems. Concepts in Magnetic Resonance Part A: An Educational Journal, 2005, 26(2): 78-90.
Luo, G., Bijeljic, B., Luo, S., et al. A lightweight multiscale neural network for inversion of NMR relaxation measurements in porous media. Transport in Porous Media, 2025, 152(4): 28.
Maletin, N. V., Eremenko, A. M., Minaev, D. V. An algorithm for solving a 1D seismic data inversion problem on quantum or digital annealers in the presence of a priori information on layer parameters specified by arbitrary functions. Laser Physics Letters, 2025, 22(1): 015201.
Wang, Y., Wen, X. Quantum computational methods for inverse problems. Journal on Numerical Methods and Computer Applications, 2025, 46(2): 95-115.
Downloads
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Jiangfeng Guo, Qiaosheng Wan, Qimeng Li, Ranhong Xie, Lizhi Xiao
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.