1: \begin{abstract}

2: In recent years, numerous vision and learning tasks have been (re)formulated as nonconvex and nonsmooth programmings (NNPs).

3: Although some algorithms have been proposed for particular problems, designing fast and flexible optimization schemes with theoretical guarantee is a challenging task for general NNPs.

4: It has been investigated that performing inexact inner iterations often benefit to special applications case by case, but their convergence behaviors are still unclear.

5: Motivated by these practical experiences, this paper designs a novel algorithmic framework, named inexact proximal alternating direction method (IPAD) for solving general NNPs.

6: We demonstrate that any numerical algorithms can be incorporated into IPAD for solving subproblems and the convergence of the resulting hybrid schemes can be consistently guaranteed by a series of simple error conditions.

7: Beyond the guarantee in theory, numerical experiments on both synthesized and real-world data further demonstrate the superiority and flexibility of our IPAD framework for practical use.

8: \end{abstract}

9: