Abstract
Deep neural network based surrogates for partial differential equations haverecently gained increased interest. However, akin to their numericalcounterparts, different techniques are used across applications, even if theunderlying dynamics of the systems are similar. A prominent example is theLagrangian and Eulerian specification in computational fluid dynamics, posing achallenge for neural networks to effectively model particle- as opposed togrid-based dynamics. We introduce Universal Physics Transformers (UPTs), anovel learning paradigm which models a wide range of spatio-temporal problems -both for Lagrangian and Eulerian discretization schemes. UPTs operate withoutgrid- or particle-based latent structures, enabling flexibility across meshesand particles. UPTs efficiently propagate dynamics in the latent space,emphasized by inverse encoding and decoding techniques. Finally, UPTs allow forqueries of the latent space representation at any point in space-time. Wedemonstrate the efficacy of UPTs in mesh-based fluid simulations, steady-stateReynolds averaged Navier-Stokes simulations, and Lagrangian-based dynamics.Project page: https://ml-jku.github.io/UPT