Operator Compression with Deep Neural Networks

Fabian Kröpfl, Roland Maier, Daniel Peterseim

This paper studies the compression of partial differential operators using neural networks. We consider a family of operators, parameterized by a potentially high-dimensional space…