Uncertainties in Physics-informed Inverse Problems: The Hidden Risk in Scientific AI
This addresses the risk of selecting non-physical solutions in scientific AI, which is crucial for researchers in physics and machine learning, though it appears incremental as it builds on existing PIML methods.
The paper tackles the problem of uncertainties in physics-informed machine learning (PIML) for inverse problems, such as estimating coefficient functions, by proposing a framework to quantify and analyze these uncertainties, and it shows that unique identification is possible with geometric constraints, as confirmed in a reduced magnetohydrodynamics model.
Physics-informed machine learning (PIML) integrates partial differential equations (PDEs) into machine learning models to solve inverse problems, such as estimating coefficient functions (e.g., the Hamiltonian function) that characterize physical systems. This framework enables data-driven understanding and prediction of complex physical phenomena. While coefficient functions in PIML are typically estimated on the basis of predictive performance, physics as a discipline does not rely solely on prediction accuracy to evaluate models. For example, Kepler's heliocentric model was favored owing to small discrepancies in planetary motion, despite its similar predictive accuracy to the geocentric model. This highlights the inherent uncertainties in data-driven model inference and the scientific importance of selecting physically meaningful solutions. In this paper, we propose a framework to quantify and analyze such uncertainties in the estimation of coefficient functions in PIML. We apply our framework to reduced model of magnetohydrodynamics and our framework shows that there are uncertainties, and unique identification is possible with geometric constraints. Finally, we confirm that we can estimate the reduced model uniquely by incorporating these constraints.