1: \begin{abstract}

2: Polynomial matrix inequalities can be solved using hierarchies of convex relaxations,

3: pioneered by Henrion and Lassere. 

4: %based on the original scalar polynomial optimization relaxations of Lassere. However, 

5: In some cases, this might not be practical, and one may need to resort to methods 

6: with local convergence guarantees, whose development has been rather \emph{ad hoc}, so far.

7: In this paper, we explore several alternative approaches to the problem, 

8: with non-trivial guarantees available using results from tame geometry. 

9: %that when the underlying geometric structure of the non-convex optimization problem has a tame nature, one can use first-order methods without the need to sacrifice the theoretical convergence guarantees. 

10: \end{abstract}

11: