[2512.07755] Physics-Informed Neural Networks for Source Inversion and Parameters Estimation in Atmospheric Dispersion

Statistics > Machine Learning arXiv:2512.07755 (stat) [Submitted on 8 Dec 2025 (v1), last revised 8 Apr 2026 (this version, v2)] Title:Physics-Informed Neural Networks for Source Inversion and Parameters Estimation in Atmospheric Dispersion Authors:Brenda Anague, Bamdad Hosseini, Issa Karambal, Jean Medard Ngnotchouye View a PDF of the paper titled Physics-Informed Neural Networks for Source Inversion and Parameters Estimation in Atmospheric Dispersion, by Brenda Anague and 3 other authors View PDF HTML (experimental) Abstract:Recent studies have shown the success of deep learning in solving forward and inverse problems in engineering and scientific computing domains, such as physics-informed neural networks (PINNs). In the fields of atmospheric science and environmental monitoring, estimating emission source locations is a central task that further relies on multiple model parameters that dictate velocity profiles and diffusion parameters. Estimating these parameters at the same time as emission sources from scarce data is a difficult task. In this work, we achieve this by leveraging the flexibility and generality of PINNs. We use a weighted adaptive method based on the neural tangent kernels to solve a source inversion problem with parameter estimation on the 2D and 3D advection-diffusion equations with unknown velocity and diffusion coefficients that may vary in space and time. Our proposed weighted adaptive method is presented as an extension of PINNs for forward PDE p...