A math problem so difficult it was considered computationally impossible can now be solved by a neural network using random rough functions.
Researchers proved that the complex H-1 mathematical norm can be recovered by averaging evaluations against random test functions. This approach bypasses the need for massive computing power that was previously required to solve these variational physics equations. Neural networks can now use this shortcut to model fluid dynamics and structural stress with extreme precision. The math community generally considered this specific norm to be a natural but unusable metric for practical simulation. This breakthrough turns a theoretical dead end into a high-speed tool for engineering and physics. It allows us to simulate the real world with a level of accuracy that was previously out of reach.
Random test functions, $H^{-1}$ norm equivalence, and stochastic variational physics-informed neural networks
arXiv · 2605.03542
The dual norm characterisation of weak solutions of second-order linear elliptic partial differential equations is mathematically natural but computationally intractable: evaluating the $H^{-1}$ norm of a residual requires a supremum over an infinite-dimensional function space. We prove that the $H^{-1}$ norm of any functional is equivalent to its expected squared evaluation against a random test function whose distribution depends only on the domain. Crucially, realisations of this random test