1: \begin{abstract}

2: We propose new restarting strategies for accelerated gradient and accelerated coordinate descent methods.

3: Our main contribution is to show that the restarted method has a geometric rate 

4: of convergence for any restarting frequency, and so it allows us to take profit of restarting even when we do not know the strong convexity coefficient.

5: The scheme can be combined with adaptive restarting, leading to the first 

6: provable convergence for adaptive restarting schemes with accelerated gradient methods.

7: Finally, we illustrate the properties of the algorithm on a regularized logistic regression problem and on a Lasso problem.

8: \end{abstract}

9: