New MiNO method learns PDE propagators for improved scientific machine learning

Researchers have developed MiNO, a novel approach for learning partial differential equations (PDEs) by focusing on the propagator rather than the solution field or map. This method utilizes the eikonal equation for phase and the transport equation for amplitude to reconstruct solutions via oscillatory integrals. MiNO demonstrates superior accuracy and efficiency compared to traditional physics-informed neural networks and supervised Fourier neural operators, particularly on benchmarks involving discontinuous advection, and can generalize to unseen initial conditions without retraining. AI

IMPACT Introduces a novel method for solving PDEs that shows improved accuracy and generalization capabilities over existing techniques.

RANK_REASON The cluster contains a research paper detailing a new method for scientific machine learning. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• eikonal equation
• MiNO
• neural operators
• oscillatory integral
• partial differential equations
• physics-informed neural networks
• supervised Fourier neural operator

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →