Partial differential equations in the age of machine learning: a critical synthesis of classical, machine learning, and hybrid methods

Abstract

Abstract Partial differential equations (PDEs) govern physical phenomena across the full range of scientific scales, yet their computational solution remains one of the defining challenges of modern science. This critical review examines two mature but epistemologically distinct paradigms for PDE solution, classical numerical methods and machine learning approaches, through a unified evaluative framework structured around six fundamental computational challenges: high dimensionality, nonlinearity, geometric complexity, discontinuities, multiscale phenomena, and multiphysics coupling. Classical methods, including finite difference, finite element, finite volume, and spectral discretizations, are assessed for their structure-preserving properties, rigorous convergence theory, and scalable solver design; their persistent limitations in high-dimensional and geometrically complex settings are characterized in detail. Machine learning approaches, including physics-informed neural networks, neural operators, graph architectures, transformers, generative models, and hybrid frameworks, are evaluated under a taxonomy organized by the degree to which physical knowledge is incorporated, and subjected to the same critical assessment applied to classical methods. The central finding is that classical methods are deductive: errors are bounded by quantities derivable from PDE structure and discretization parameters. Machine learning methods are inductive: accuracy depends on statistical proximity to the training distribution. This distinction determines not which paradigm is superior, but what each can certify, and therefore where each should be deployed. Epistemological character, rather than computational speed, is the primary criterion governing responsible method selection. We identify three genuine complementarities between the paradigms and develop principles for hybrid design, including a framework for the structure inheritance problem and an error budget decomposition that separates discretization, neural approximation, and coupling contributions. We further assess emerging frontiers, including foundation models for scientific computing, differentiable programming for inverse design, quantum algorithms, and exascale co-design, evaluating each against the structural constraints that determine whether current barriers are fundamental or contingent on engineering progress. The conclusion is not that one paradigm supersedes the other, but that their principled integration, grounded in the complementary strengths each provides, offers the most credible path toward computationally tractable solutions to the most consequential open problems that neither can address alone.