1: \begin{abstract}

2: We prove that supercritical branching random walk on a transient graph converges almost surely under rescaling to a random measure on the Martin boundary of the graph. Several open problems and conjectures about this limiting measure are presented.

3: % Let $G$ denote the Cayley graph of a non-amenable finitely generated group, and consider a supercritical (i.e., with positive chance of indefinite survival) and transient branching random walk (BRW) on $G$.

4: % Such BRW can be seen as a process on $G$, as well as a Markov chain on the set of its possible configurations.

5: % First we show that a ``renormalized'' version of BRW induces a \emph{random measure} on the Martin boundary for the underlying random walk.

6: % Subsequently, we analyze an example where such random measure generates the tail $\sigma$-algebra of the BRW seen as a Markov chain on the set of its configurations.

7: % We believe this to be a general fact, which we state in Conjecture \ref{conj:sigma}.

8: % %

9: % %

10: % %-----

11: % %

12: % %, we investigate the relations between these two points of view.

13: % %We show that in the first interpretation BRW induces a \emph{random} measure on the Martin boundary for the underlying random walk.

14: % %Meanwhile, classical Martin boundary theory guarantees that (in the second interpretation) BRW seen as a Markov chain induces a limiting measure on the Martin boundary of the configuration graph.

15: % %We investigate the connections between such limiting measures and conjecture that they induce the same tail sigma-algebra, we also present an example where the conjecture holds.

16: % To our knowledge, this is the first work comparing different Martin boundaries investigating how they are related to each other.

17: \end{abstract}

18: