1: \begin{abstract}

2:   This paper considers continuously differentiable functions of two

3:   vector variables that have (possibly a continuum of) min-max saddle

4:   points. We study the asymptotic convergence properties of the

5:   associated saddle-point dynamics (gradient-descent in the first

6:   variable and gradient-ascent in the second one).  We identify a

7:   suite of complementary conditions under which the set of saddle

8:   points is asymptotically stable under the saddle-point dynamics.

9:   Our first set of results is based on the convexity-concavity of the

10:   function defining the saddle-point dynamics to establish the

11:   convergence guarantees.  For functions that do not enjoy this

12:   feature, our second set of results relies on properties of the

13:   linearization of the dynamics, the function along the proximal

14:   normals to the saddle set, and the linearity of the function in one

15:   variable.  We also provide global versions of the asymptotic

16:   convergence results.  Various examples illustrate our discussion.

17: \end{abstract}

18: