1: \begin{proof}

2: Applying Theorem~\ref{thm:main_result3} to the pseudo-Lipschitz loss function $\psi^1: \R^{\p} \times \R^{\p} \rightarrow \R$, defined as $\psi^1(\mathbf{x}, \mathbf{y}) = ||\mathbf{x}-\mathbf{y}||^2/\p$, we find

3: %

4: $\plim_{\p} \frac{1}{\p}\norm{ \widehat{ \bet}- \bet}^2  = \lim_t \plim_{\p} \frac{1}{\p} \E_{\mathbf{Z}}[\norm{  \prox_{J_{\bfalph \tau_t}}(\bet + \tau_t \mathbf{Z})- \bet}^2].$  

5: %

6: The desired result follows since $\lim_t \plim_{\p} \frac{1}{\p} \E_{\mathbf{Z}}[\norm{  \prox_{J_{\bfalph \tau_t}}(\bet + \tau_t \mathbf{Z})- \bet}^2] = \delta(\tau_{*}^2 - \sigma_w^2)$.  To see this, note that  $\lim_t \delta(\tau_{t+1}^2 - \sigma_w^2)  = \delta(\tau_{*}^2 - \sigma_w^2) $ and %

7: \[\plim_{\p} \frac{1}{\p} \E_{\mathbf{Z}}[\norm{  \prox_{J_{\bfalph \tau_t}}(\bet + \tau_t \mathbf{Z})- \bet}^2] = \lim_{\p}  \frac{1}{\p}\E_{\mathbf{Z}, \bm B}[  \norm{  \prox_{J_{\bfalph \tau_t}}(\mathbf{B} + \tau_t \mathbf{Z})- \mathbf{B}}^2] = \delta(\tau_{t+1}^2 - \sigma_w^2),\]

8: %

9: for $\mathbf{B}$ elementwise i.i.d.\ $B$ independent of  $\mathbf{Z}  \sim \mathcal{N}(0, \mathbb{I}_{\p})$.  A rigorous argument for the above requires showing that the assumptions of Lemma~\ref{lem:beta_expectation} are satisfied and follows similarly to that used to prove property \textbf{(P2)} stated in Section \ref{sec_mainproof} and proved in Appendix~\ref{app_props}.

10: \end{proof}

11: