1: \begin{abstract}

2:   We study variational regularization methods in a general framework,

3:   more precisely those methods that use a discrepancy and a regularization

4:   functional. While several sets of sufficient conditions are known to

5:   obtain a regularization method, we start with an investigation of

6:   the converse question: How could necessary conditions for a

7:   variational method to provide a regularization method look like? To this end,

8:   we formalize the notion of a variational scheme and start with comparison

9:   of  three

10:   different instances of variational methods. Then we focus on the

11:   data space model and investigate the role and interplay of the

12:   topological structure, the convergence notion and the discrepancy

13:   functional. Especially, we deduce necessary conditions for the

14:   discrepancy functional to fulfill usual continuity assumptions.

15:   The results are applied to discrepancy functionals given by Bregman

16:   distances and especially to the Kullback-Leibler divergence. 

17: \end{abstract}

18: