1 paper · 1 filter
Patrick Gelß, Aizhan Issagali, Ralf Kornhuber
We present a novel and mathematically transparent approach to function approximation and the training of large, high-dimensional neural networks, based on the approximate least-squ…