1: \begin{abstract}

2: %% Abstract must have 150-250 words. Our abstract has 172 (not counting with keywords).

3: In this work, we propose an algorithm, StochaLM, which finds the solution of an optimization problem by solving a sequence of subproblems where some components of the cost are linearized around specific anchor points. Our method converges to the minimizer of the objective with probability one, and under very general assumptions. Namely, our method can tackle problems with convex constraints embedded in the cost, and requires strong convexity only in one component of the objective. Our assumptions on the stochasticity of the method are more general than the assumptions in related methods that exist in the literature. Under these more general assumptions, the proofs of convergence of these related methods become invalid; the proof of convergence of our method is, therefore, also novel. Our method is application friendly, in the sense that it has no hyperparameters to tune. Additionally, our method is easily adaptable to both federated learning and decentralized optimization setups. Numerical experiments evidence that our method is competitive with state-of-the-art methods in both centralized and distributed applications.

4: \end{abstract}

5: