1: \begin{abstract}

2: We propose and analyze the convergence of a novel stochastic forward-backward splitting algorithm 

3:  for solving monotone inclusions given by the sum of a maximal monotone

4: operator and a single-valued maximal monotone cocoercive operator. This latter framework

5: has a number of interesting special cases, including variational inequalities and convex minimization

6: problems,   while  stochastic approaches are practically relevant to account for perturbations in the data.     

7: The algorithm we propose  is a stochastic extension of  the classical deterministic forward-backward method, and is obtained  considering  the composition of the resolvent of the maximal monotone operator with  

8: a forward step based on a stochastic estimate of the single-valued operator. 

9: Our study  provides a non-asymptotic error analysis  in expectation for the strongly 

10: monotone case, as well as  almost sure convergence   under weaker assumptions. 

11: The approach we consider allows to avoid averaging, a feature critical when considering methods based on sparsity, and,  for  minimization problems, it allows  to obtain convergence rates  matching  those obtained by stochastic extensions of so called accelerated methods.

12: Stochastic quasi Fejer's sequences are  a key technical  tool  to prove almost sure convergence. 

13: \end{abstract}

14: