1: \begin{abstract}~

2:   In this paper, we provide bounds in Wasserstein and total variation distances between the distributions of

3:    the successive iterates of two functional autoregressive processes

4:    with isotropic Gaussian noise of the form

5:    $Y_{k+1} = \rmT_\gamma(Y_k) + \sqrt{\gamma\sigma^2} Z_{k+1}$ and

6:    $\tilde{Y}_{k+1} = \tilde{\rmT}_\gamma(\tilde{Y}_k) +

7:    \sqrt{\gamma\sigma^2} \tilde{Z}_{k+1}$.

8:    % in the limit where the

9:    % parameter $\gamma \to 0$.

10:    More precisely, we give   non-asymptotic bounds on

11:    $\rho(\mathcal{L}(Y_{k}),\mathcal{L}(\tilde{Y}_k))$, where $\rho$ is

12:    an appropriate weighted Wasserstein distance or a $V$-distance,

13:    uniformly in the parameter $\gamma$, and on $\rho(\pi_{\gamma},\tilde{\pi}_{\gamma})$,

14:    where  $\pi_{\gamma}$ and $\tilde{\pi}_{\gamma}$  are the respective

15:     stationary   measures of the two processes. The class of considered processes encompasses the

16:    Euler-Maruyama discretization of Langevin diffusions and its

17:    variants.  The bounds we derive are of order $\gamma$ as $\gamma \to 0$. To obtain

18:    our results, we  rely on the construction of a discrete sticky Markov chain

19:    $(W_k^{(\gamma)})_{k \in \nset}$ which bounds  the distance between an appropriate coupling of the two processes.  We then establish stability and quantitative convergence results for this process uniformly on

20:    $\gamma$. In addition, we show that it converges in

21:    distribution to the continuous sticky process studied in

22:    \cite{howitt2007stochastic,eberle:zimmer:2016}. Finally, we

23:    apply our result to Bayesian inference of ODE parameters and  numerically illustrate them on two particular problems.

24:  \end{abstract}

25: