1: \begin{abstract}

2: We propose a globally convergent multilevel training method for deep residual networks (ResNets). 

3: The devised method can be seen as a novel variant of the recursive multilevel trust-region (RMTR) method, which operates in hybrid (stochastic-deterministic) settings by adaptively adjusting mini-batch sizes during the training. 

4: The multilevel hierarchy and the transfer operators are constructed by exploiting a dynamical system's viewpoint, which interprets forward propagation through the ResNet as a forward Euler discretization of an initial value problem. 

5: In contrast to traditional training approaches, our novel RMTR method also incorporates curvature information on all levels of the multilevel hierarchy by means of the limited-memory SR1 method.

6: The overall performance and the convergence properties of our  multilevel training method are numerically investigated using examples from the field of classification and regression.

7: \end{abstract}

8: