1: \begin{abstract}

2: 

3: Recently, decentralized optimization over the Stiefel manifold

4: has attacked tremendous attentions 

5: due to its wide range of applications in various fields.

6: Existing methods rely on the gradients to update variables,

7: which are not applicable to the objective functions

8: with non-smooth regularizers,

9: such as sparse PCA.

10: In this paper, to the best of our knowledge,

11: we propose the first decentralized algorithm 

12: for non-smooth optimization over Stiefel manifolds.

13: Our algorithm approximates the non-smooth part of objective function 

14: by its Moreau envelope,

15: and then existing algorithms for smooth optimization can be deployed.

16: We establish the convergence guarantee 

17: with the iteration complexity of $\cO (\epsilon^{-4})$.

18: Numerical experiments conducted under the decentralized setting

19: demonstrate the effectiveness and efficiency of our algorithm.

20: 

21: 

22: \end{abstract}

23: