1: \begin{abstract}

2: We propose and study a weakly convergent variant of the forward--backward algorithm for solving structured monotone inclusion problems.

3: Our algorithm features a per-iteration deviation vector which provides additional degrees of freedom.

4: The only requirement on the deviation vector to guarantee convergence is that its norm is bounded by a quantity that can be computed online.

5: This approach provides great flexibility and opens up for the design of new and improved forward--backward-based algorithms, while retaining global convergence guarantees.

6: These guarantees include linear convergence of our method under a metric subregularity assumption without the need to adapt the algorithm parameters.

7: 

8: Choosing suitable monotone operators allows for incorporating deviations into other algorithms, such as Chambolle--Pock and \Krasnoselsky--Mann iterations.

9: We propose a novel inertial primal--dual algorithm by selecting the deviations along a momentum direction and deciding their size using the norm condition.

10: Numerical experiments demonstrate our convergence claims and show that even this simple choice of deviation vector can improve the performance, compared, e.g., to the standard Chambolle--Pock algorithm.

11: \end{abstract}

12: