1: \begin{abstract}

2: We study the convergence of variationally regularized solutions to linear ill-posed operator 

3: equations in Banach spaces as the noise in the right hand side tends to $0$. The rate of 

4: this convergence is determined by abstract smoothness conditions on the solution called 

5: source conditions. For non-quadratic data fidelity or penalty terms such source conditions are 

6: often formulated in the form of variational inequalities. Such variational source conditions (VSCs) 

7: as well as other formulations of such conditions in Banach spaces have the disadvantage 

8: of yielding only low-order convergence rates. A first step towards higher order 

9: VSCs has been taken by Grasmair (2013) who obtained convergence 

10: rates up to the saturation of Tikhonov regularization. For even higher order convergence rates, 

11: iterated versions of variational regularization have to be considered. In this paper we 

12: introduce VSCs of arbitrarily high order which lead to optimal convergence 

13: rates in Hilbert spaces. For Bregman iterated variational regularization in Banach spaces with general data fidelity 

14: and penalty terms, we derive convergence rates under third order VSC. 

15: These results are further discussed for entropy regularization with elliptic pseudodifferential 

16: operators where the VSCs are interpreted in terms of Besov spaces and the optimality of the rates 

17: can be demonstrated. Our theoretical results are confirmed in numerical experiments. 

18: 

19: \keywords{inverse problems, variational regularization, convergence rates, entropy regularization, 

20: variational source conditions}

21: %{65J20 \and 65J22}

22: \end{abstract}

23: