Johannes Brandstetter

Generative models can now capture the full spectrum of solutions in high-dimensional physical systems with symmetry-breaking bifurcations, a feat previously intractable.

Achieving zero-variance estimators in stochastic systems by eliminating non-adiabatic lag could revolutionize free energy estimation techniques.

GyroFlow achieves rapid and accurate generation of steady-state turbulence statistics without the need for resolving transient dynamics, revolutionizing how we approach gyrokinetic simulations.

Achieve up to 7% performance gains in out-of-distribution simulation tasks by adapting high-dimensional surrogates at test time using D-optimal statistics.