1: \begin{proof}

2: First, let us rewrite the ASP algorithm in a AMP-like form as

3: considered in \cite{DeshpandeM14,deshpande2015asymptotic}

4: \begin{align}

5: {\bf T}^t =& \frac{1}{\sqrt N} S \hat {\bf x}^t - b^t \hat {\bf

6:              x}^{t-1}\, , \label{flo2-1} \\

7: \hat x_{i}^{t+1} =& \eta_t({ T}_i^t)\, , \label{flo2-2}

8: \end{align}

9: with, again 

10: \begin{align}

11: b^t = \frac{ {\mathbb E} [S^2] }N

12:   \sum_{i=1}^N(\partial_T\eta_t({T}_i^t))\, .  \label{flo2-3}

13: \end{align}

14: %

15: With the particular choice of the following denoising function

16: \begin{align}

17: \eta_t({ T}_i^t) \eqdef \frac1s \frac{\partial}{\partial

18:   T_{i}}f_{\text{in}}\left[T_i^t,V^{(1),t},V^{(0),t}\right]  \label{flo2-4}

19: \end{align}

20: where $V^{(1),t}$, $V^{(1),t}$ are deterministic variables given by

21: eqs.~(\ref{eq:def_V1_SE}-\ref{eq:def_V0_SE}) with use of eqs.~(\ref{eq:1RSB_SE_Delta0}-\ref{eq:1RSB_SE_Delta1}).

22: %\begin{align}

23: %V^{(1),t} &= \frac 1\D \left(\D^{(1),t}+\D^{(0),t}+Q^t            \right)-\frac{\D_0}{\D^2}\D^{(1),t} \, ,& V^{(0),t}&=\frac{\D_0}{\D^2}\D^{(0),t}

24: %\end{align}

25: %and

26: %\begin{align}

27: %\D^{(0),t} &= {\mathbb{E}}[\D^{(0),t} (Z) ]  \, , &

28: %\D^{(1),t} &={\mathbb{E}}[\D^{(1),t} (Z) ] 

29: %\end{align}

30: %{\bf it would be useful to defined the corresponding function here,

31: %  otherwise it is hard to formally state the theorem} 

32: Once ASP is written in this form, we can apply

33: directly theorem \ref{theorem1} from section \ref{sec:rigorous_RS} and reach the desired result. 

34: \end{proof}

35: