arXiv：cs.LG· Jae Deok Kim, Sai Ravela, Rob. L. Evans·· 3 小时前AI 评分31
神经物理反演器 NPI：集合条件化与残差学习耦合的模块化大地电磁反演框架
The Neuro-Physical Inverter: A Modular Framework for Magnetotelluric Inversion Coupling Ensemble Conditioning with Residual Learning
AI 导读
研究者提出神经物理反演器（NPI），一个将集合条件化与约束残差学习耦合的模块化、不确定性感知地球物理反演框架，以一维大地电磁（MT）为受控试验场景。NPI 先用集合条件高斯过程（EnsCGP）生成物理可行的参考集合，再由残差学习网络预测修正，并在美国内华达州 Gabbs Valley 地热区宽带 MT 数据上降低了中周期频段的跨台站平均失配，同时保持相当的集合离散度。
正文
Abstract:We present the Neuro-Physical Inverter (NPI), a modular, uncertainty-aware framework for geophysical inversion that couples ensemble-based conditioning with constrained residual learning, demonstrated in the 1D magnetotelluric (MT) setting as a controlled testbed. The framework operates in two stages. An Ensemble-Conditional Gaussian Process (EnsCGP) conditions a prior ensemble of resistivity models on the observed response, producing a physically admissible reference ensemble. A residual-learning neural network then predicts targeted corrections to this reference, trained on synthetic data and fine-tuned per station for field application through a physics-coupled objective. Because an ensemble is conditioned, refined, and propagated through both stages, every estimate carries an associated ensemble spread. Synthetic experiments show that NPI systematically reduces ensemble-mean error without destabilizing the ensemble. Applied to broadband MT data from the Gabbs Valley geothermal region (Nevada, USA), NPI reduces the across-station mean misfit over the mid-period band while retaining comparable ensemble spread. The propagated ensemble yields a factor of uncertainty that serves as an operational measure of constraint within the assumed model class. Both stages are dimension-agnostic in formulation, and the design principles established here are intended to scale to higher-dimensional parameterizations.
| Comments: | Accepted by IEEE TGRS |
| Subjects: | Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.03225 [cs.LG] |
| (or arXiv:2610.03225v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.03225 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Jae Deok Kim [view email]
[v1]
Fri, 2 Oct 2026 12:41:43 UTC (20,632 KB)
来源：arXiv：cs.LG · arxiv.org