I read the paper "link prediction based on graph neural networks", In Theorem 2 of this paper, it use: $$\eta \sum_{l=a h+b+1}^{\infty} \gamma^{l} \lambda^{l}=\eta(\gamma \lambda)^{a h+b+1}(1-\gamma \lambda)^{-1}$$
where $\gamma$ is a decaying factor between 0 and 1, $\lambda < \frac{1}{\gamma}$. $\eta$ is a positive constant or a positive function of $\gamma$ that is upper bounded by a constant, $a, b \in \mathbb{N}$ and $a > 0$, $h$ is a positive integer. I can not understand why the equation holds.