High-Dimensional Distribution Generation Through Deep Neural Networks

no code implementations • 26 Jul 2021 • Dmytro Perekrestenko, Léandre Eberhard, Helmut Bölcskei

We show that every $d$-dimensional probability distribution of bounded support can be generated through deep ReLU networks out of a $1$-dimensional uniform input distribution.

Quantization Vocal Bursts Intensity Prediction