Picture for Florian Rossmannek

Florian Rossmannek

Fading memory and the convolution theorem

Add code
Aug 14, 2024
Viaarxiv icon

State-Space Systems as Dynamic Generative Models

Add code
Apr 12, 2024
Viaarxiv icon

Gradient descent provably escapes saddle points in the training of shallow ReLU networks

Add code
Aug 03, 2022
Viaarxiv icon

Landscape analysis for shallow ReLU neural networks: complete classification of critical points for affine target functions

Add code
Mar 19, 2021
Figure 1 for Landscape analysis for shallow ReLU neural networks: complete classification of critical points for affine target functions
Figure 2 for Landscape analysis for shallow ReLU neural networks: complete classification of critical points for affine target functions
Viaarxiv icon

A proof of convergence for gradient descent in the training of artificial neural networks for constant target functions

Add code
Feb 19, 2021
Viaarxiv icon

Non-convergence of stochastic gradient descent in the training of deep neural networks

Add code
Jun 12, 2020
Viaarxiv icon