EF1 – Extracting dynamical Laws from Complex Data

Project

EF1-13

Stochastic and rough aspects in deep neural networks

Project Heads

Christian Bayer, Peter-Karl Friz

Project Members

Nikolas Tapia

Project Duration

01.01.2021 − 31.12.2022

Located at

WIAS / TU Berlin

Description

We analyse residual neural networks using rough path theory. Extending worst-case stability analysis developed in the first funding period, we now embrace the stochastic nature (random initialization, stochastic optimization) of training networks, seen as ultimate justification of rough path analysis for deep networks.

External Website

Related Publications

Related Pictures

Related Pictures