1: \begin{abstract} Consider a weighted or unweighted $k$-nearest

2:   neighbor graph that has been built on $n$ data points drawn randomly

3:   according to some density $p$ on $\R^d$. We study the convergence of

4:   the shortest path distance in such graphs as the sample size 

5:   tends to infinity.  We prove that for unweighted $\knn$ graphs, this

6:   distance converges to an unpleasant distance function on the

7:   underlying space whose properties are detrimental to machine

8:   learning. We also study the behavior of the shortest path distance in weighted

9:   $\knn$ graphs.  

10: \end{abstract}