Team Director, RIKEN
1 paper at NeurIPS 2025
We propose a method for learning energy-preserving or -dissipating numerical schemes. Our approach admits the universal approximation property.
