presentation-colbisbm/principal.tex

554 lines
No EOL
23 KiB
TeX

\section{Model Context}
\label{sec:context-of-the-model}
\begin{frame}
\frametitle{Why a network?}
\begin{columns}
\begin{column}{0.5\textwidth}
\begin{columns}
\begin{column}{0.5\textwidth}
\begin{figure}[ht]
\centering
\begin{tikzpicture}[scale=.6,rotate=270]
\input{tikz/plantpollinatornetwork.tex}
\end{tikzpicture}
\caption{Example of a network}
\label{fig:plants-pollin}
\end{figure}
\end{column}
\begin{column}{0.3\textwidth}
\centering
\begin{align*}
\begin{pmatrix}
1 & 0 & 1 \\
1 & 0 & 0 \\
1 & 0 & 0 \\
1 & 1 & 0
\end{pmatrix}
\end{align*}
\footnotesize
Associated adjacency matrix
\end{column}
\end{columns}
\begin{figure}[ht]
\centering
\includegraphics[width=0.7\textwidth]{tikz/applications/baldock/graph-Baldock2019_Bristol.pdf}
\caption{Plant-pollinator network of
Bristol\newline\cite{baldockSystemsApproachReveals2019}}
\label{fig:label}
\end{figure}
\end{column}
\begin{column}{0.5\textwidth}
\begin{itemize}
\item Modeling of various interactions, here ecosystems
\item Structure necessary for: biodiversity monitoring, robustness, risk
of collapse
\item Increasingly available
\end{itemize}
\end{column}
\end{columns}
\end{frame}
\begin{frame}{Analysis methods for a network}
Several methods~:
\begin{itemize}
\item Metrics~: degree, centrality, nesting \dots
\item Network embedding with GNN
\item \textbf<2>{\emph{Clustering} of nodes with latent variable models}
\end{itemize}
\end{frame}
\begin{frame}
\addtocounter{footnote}{1}
\frametitle{Latent Block Model (LBM\footnotemark[\thefootnote])}
%DONE remplacer i \in bullet par Zi = \bullet
\cite{govaertEMAlgorithmBlock2005}.
\begin{columns}
\begin{column}{0.40\linewidth}
\begin{figure}[H]
\center
\begin{tikzpicture}[scale=0.35]
\input{tikz/lbm.tex}
\end{tikzpicture}
\caption{Example of LBM\footnotemark[\thefootnote]}
\label{fig:LBMvisu}
\end{figure}
\end{column}
\only<1>{
\begin{column}{0.51\linewidth}
\begin{block}{Hierarchical model}
\vspace{-\baselineskip}
\begin{align*}
\forall q\in[\![ 1, Q_1]\!],~ & \mathbb{P}(Z_i = q) = \pi_q \\
\forall r\in[\![ 1, Q_2]\!],~ & \mathbb{P}(W_j = r) = \rho_r \\
& Y_{ij} | Z_i, W_j \sim \mathcal{F}(\alpha_{Z_i,W_j})
\end{align*}
where $|\pi| = Q_1, |\rho| = Q_2, |\alpha| = Q_1 \times Q_2$
\end{block}
\begin{block}{Concise LBM formula}
$Y \sim \mathcal{F}\text{-BiSBM}_{n_1,n_2}(Q_1, Q_2, \pi, \rho, \alpha)$
\end{block}
\end{column}}
\only<2>{
\begin{column}{0.51\linewidth}
With \begin{itemize}
\item $Q_1 = |\{{\color{blueind}\bullet},{\color{cyanind}\bullet},{\color{electricblue}\bullet}\}|$ fixed row blocks
\item $Q_2 = |\{{\color{burntorange}\bullet},{\color{goldenyellow}\bullet},{\color{peach}\bullet}\}|$ fixed column blocks
\end{itemize}
\begin{block}{Parameters}
\begin{itemize}
\item $\pi_{{\color{blueind}\bullet}} = \mathbb{P}(Z_i = {\color{blueind}\bullet})$
\item $\rho_{{\color{burntorange}\bullet}} = \mathbb{P}(W_j = {\color{burntorange}\bullet})$
\item $\alpha_{{\color{blueind}\bullet}{\color{burntorange}\bullet}} = \mathbb{P}(Y_{ij} = 1 | Z_i = {\color{blueind}\bullet}, W_j = {\color{burntorange}\bullet})$
\end{itemize}
\end{block}
\end{column}}
\end{columns}
\footnotetext[\thefootnote]{Which I will henceforth call BiSBM}
\end{frame}
\begin{frame}
\frametitle{Multiple networks}
\begin{figure}[ht]
\centering
\begin{subfigure}[ht]{0.3\textwidth}
\includegraphics[width=1.1\textwidth]{tikz/applications/baldock/mat-Baldock2019_Bristol.pdf}
\caption{Bristol}
\end{subfigure}
\begin{subfigure}[ht]{0.3\textwidth}
\includegraphics[width=1.1\textwidth]{tikz/applications/baldock/mat-Baldock2019_Edinburgh.pdf}
\caption{Edinburgh}
\end{subfigure}
\begin{subfigure}[ht]{0.3\textwidth}
\includegraphics[width=1.1\textwidth]{tikz/applications/baldock/mat-Baldock2019_Leeds.pdf}
\caption{Leeds}
\end{subfigure}
\caption{Adjacency matrices,~\cite{baldockSystemsApproachReveals2019}}
\label{fig:adj}
\end{figure}
\end{frame}
\section[Bipartite collection models]{Bipartite network collection models}
\label{sec:extension-of-colsbm-to-bipartite-networks}
\begin{frame}
\frametitle{Bipartite collections}
\[
\forall m \in \{1\dots M\}, Y^m \overset{ind}{\sim} \mathcal{F}\text{-BiSBM}_{n_1^m,n_2^m}(Q_1^m, Q_2^m, \pi^m, \rho^m, \alpha^m)
\]
\onslide<2>{
\begin{figure}[ht]
\centering
\begin{subfigure}[ht]{0.3\textwidth}
\includegraphics[width=1.1\textwidth]{tikz/applications/baldock/bisbm-mat-Baldock2019_Bristol.pdf}
\caption{Bristol}
\end{subfigure}
\begin{subfigure}[ht]{0.3\textwidth}
\includegraphics[width=1.1\textwidth]{tikz/applications/baldock/bisbm-mat-Baldock2019_Edinburgh.pdf}
\caption{Edinburgh}
\end{subfigure}
\begin{subfigure}[ht]{0.3\textwidth}
\includegraphics[width=1.1\textwidth]{tikz/applications/baldock/bisbm-mat-Baldock2019_Leeds.pdf}
\caption{Leeds}
\end{subfigure}
\caption{Reordered adjacency matrices, thanks to LBM}
\label{fig:adj-reord}
\end{figure}
}
\end{frame}
\begin{frame}
\frametitle{Different models}
\onslide<1->{ \begin{block}{\emph{iid}-colBiSBM}
\[
\forall m \in \{1\dots M\}, Y^m \overset{iid}{\sim}
\mathcal{F}\text{-BiSBM}_{n_1^m,n_2^m}(Q_1, Q_2, \pi, \rho, \alpha)
\]
with $\theta = (\pi, \rho, \alpha)$.
\end{block}}
\onslide<2>{ \begin{block}{$\pi\rho$-colBiSBM}
\[
\forall m \in \{1\dots M\}, Y^m \overset{ind}{\sim}
\mathcal{F}\text{-BiSBM}_{n_1^m,n_2^m}(Q_1, Q_2, \pi^m, \rho^m, \alpha)
\]
with $\theta = ((\pi^m)_{m=1,\dots, M}, (\rho^m)_{m=1,\dots, M}, \alpha)$.
\end{block}
}
\end{frame}
\begin{frame}
\frametitle{Parameter estimation}
% DONE say that tau i q m c' is the probability that Zim = q, approximation of the variational probability. Because we impose independence
% By maximizing a variational lower bound of the
% log-likelihood of the observed data.
Maximizing the log-likelihood?
\begin{block}{log-likelihood and complete log-likelihood}
\[
\ell(\bm{Y};\theta) = \sum_{\bm{Z,W}\in \bm{\mathcal{Z}\times\mathcal{W}}} \ell_c(\bm{Y}, \bm{Z}, \bm{W};\theta)
\]
with $\bm{\mathcal{Z}} = \{1,\dots,\alert<2>{Q_1}\}^{\alert<2>{n}},
\bm{\mathcal{W}} = \{1,\dots,\alert<2>{Q_2}\}^{\alert<2>{n}}$
\end{block}
\uncover<3>{So, classic algorithm $\Rightarrow$
\emph{Expectation-Maximization} (EM).}
\end{frame}
\begin{frame}
\frametitle{By classic EM}
At iteration $(t)$:
\begin{itemize}
\item[$\bullet$]\textbf{E Step}: calculate
$$ \mathcal{Q}(\theta | \theta^{(t-1)}) = \mathbb E_{\alert<2>{\bm Z, \bm W | \bm Y, \theta^{(t-1)}} } \left[\ell_c(\bm Y, \bm W, \bm Z; \theta) \right] $$
\item[$\bullet$]\textbf{M Step}:
$$ \theta^{(t)} = \arg \max_{\theta} \mathcal{Q}(\theta | \theta^{(t-1)})$$
\end{itemize}
\uncover<2>{
\begin{alertblock}{Problem for classic EM}
Law of $\bm{Z,W|Y},\theta^{(t-1)}$ inaccessible
\end{alertblock}}
\end{frame}
\begin{frame}
By \emph{Variational EM}, as proposed
by~\cite{daudinMixtureModelRandom2008,
chabert-liddellLearningCommonStructures2024}.
\begin{block}{Variational approximation of $\bm{Z,W|Y},\theta^{(t-1)}$}
$\mathcal{R}_{Y^m,\tau}(\mathbf{Z}^m, \mathbf{W}^m) =
\mathcal{R}^1_{Y^m,\tau}(\mathbf{Z}^m)
{\color{red}\times}
\mathcal{R}^2_{Y^m,\tau}(\mathbf{W}^m) \Rightarrow$ independence rows, columns.
\end{block}
\begin{multline*}
\ell (\bm{Y};\theta) \geq \color{red}\sum_{m=1}^{M} \bigg(
\color{black} \mathcal{Q}^m(\theta\mid\theta^{(t)}) +
\mathcal{H}(\mathcal{R}_{Y^m,\theta^{(t)}}
(\mathbf{Z}^m, \mathbf{W}^m))
\color{red}\bigg) \color{black}
\eqcolon \mathcal{J}(\tau;\theta)
\end{multline*}
where $\mathcal{Q}^m(\theta\mid\theta^{(t)}) =
\mathbb{E}_{\mathbf{Z}^m,\mathbf{W}^m
\sim \mathcal{R}_{Y^m,\tau}(.)}
\left[ \ell_c(Y^m,\mathbf{Z}^m,\mathbf{W}^m | \theta) \right] \,$
\end{frame}
\begin{frame}{Developed formula of variational EM}
\begin{multline*}
\ell (\bm{Y};\theta) \geq \color{red}\sum_{m=1}^{M} \bigg( \color{black} \sum_{i = 1}^{n_1^m}\sum_{j=1}^{n_2^m}\sum_{q \in \mathcal{Q}_{1,m}} \sum_{r \in \mathcal{Q}_{2,m}} \tau^{1,m}_{i,q} \tau^{2,m}_{j,r} \log f(Y^{m}_{ij}; \alpha_{qr}) \\
+ \sum_{i=1}^{n_1^m} \sum_{q \in \mathcal{Q}_{1,m}} \tau^{1,m}_{i,q} \log \pi_{\color{black}q}^{\color{gray}m} + \sum_{j=1}^{n_2^m} \sum_{r \in \mathcal{Q}_{2,m}} \tau^{2,m}_{j,r} \log \rho_{\color{black}r}^{\color{gray}m} \\
- \sum_{i=1}^{n_1} \tau^{1,m}_{i,q} \log \tau^{1,m}_{i,q} - \sum_{j=1}^{n_2} \tau^{2,m}_{j,r} \log \tau^{2,m}_{j,r} \color{red}\bigg) \color{black} \eqcolon
\mathcal{J}(\tau;\theta),
\end{multline*}
\begin{block}{Variational approximation}
$\tau_{iq}^{1,m} = \mathcal{R}^1_{Y^m,\tau}(Z_{iq}^m = 1)$
and $\tau_{jr}^{2,m} = \mathcal{R}^2_{Y^m,\tau}(W_{jr}^m = 1)$
\end{block}
\end{frame}
\begin{frame}{\emph{Variational Expectation} Step}
\[
\widehat{\tau}^{(t+1)} = \arg \max_{\tau}
\mathcal{J}(\mathcal{\tau},\bm{\widehat{\theta}}^{(t)})
\Leftrightarrow \arg\min_{\tau\in\mathcal{T}} \mathbf{KL}[\mathcal{R}_{\mathbf{Y},\tau}, \mathbb{P}(.|\mathbf{Y})]
\]
\begin{equation*}
\begin{cases}
\widehat{\tau}_{iq}^{1,m} \propto \widehat{\pi}_{q}^{m(t)} \prod_{j=1}^{n_2^m}\prod_{r\in\mathcal{Q}_2^m} f(Y_{ij}^m;\widehat{\alpha}_{qr}^{(t)})^{\widehat{\tau}_{jr}^{2,m(t+1)}} & \forall i = 1, \dots , n_1^m, q \in \mathcal{Q}_1^m \\
\widehat{\tau}_{jr}^{2,m} \propto \widehat{\rho}_{r}^{m(t)} \prod_{i=1}^{n_1^m}\prod_{q\in\mathcal{Q}_1^m} f(Y_{ij}^m;\widehat{\alpha}_{qr}^{(t)})^{\widehat{\tau}_{iq}^{1,m(t+1)}} & \forall j = 1, \dots , n_2^m, r \in \mathcal{Q}_2^m
\end{cases}
\end{equation*}
\footnotetext[2]{Initialization of $\widehat{\tau}$ with a
\emph{spectral clustering} on the networks.}
\end{frame}
\begin{frame}{\emph{Maximization} Step}
\[
\widehat{\theta}^{(t+1)} = \arg \max_{\theta} \mathcal{J}(\mathcal{\bm{\widehat{\tau}}}^{(t+1)},\theta)
\]
\begin{block}{Connectivity parameters}
\begin{align*}
\widehat{\alpha}_{qr} = \frac{\sum_{m=1}^{M} \sum_{i=1}^{n_1^m} \sum_{j=1}^{n_2^m} \tau_{iq}^{1,m} \tau_{jr}^{2,m} \alert<2>{Y_{ij}^m}}{\sum_{m=1}^{M} \sum_{i=1}^{n_1^m} \sum_{j=1}^{n_2^m} \tau_{iq}^{1,m} \tau_{jr}^{2,m}}
\end{align*}
\end{block}
\only<1>{
\begin{block}{Proportions for \emph{iid}}
\begin{align*}
\widehat{\pi}_q = \frac{\sum_{m=1}^{M} \sum_{i=1}^{n_1^m} \tau_{iq}^{1,m}}{\sum_{m=1}^{M} n_1^m} & &
\widehat{\rho}_r = \frac{\sum_{m=1}^{M} \sum_{j=1}^{n_2^m} \tau_{jr}^{2,m}}{\sum_{m=1}^{M} n_2^m}
\end{align*}
\end{block}
}
\only<2>{
\begin{block}{Proportions for $\pi\rho$}
\begin{align*}
\widehat{\pi}^{\color{red}m}_q = \frac{\sum_{i=1}^{n_1^m} \tau_{iq}^{1,m}}{n_1^m} & &
\widehat{\rho}^{\color{red}m}_r = \frac{\sum_{j=1}^{n_2^m} \tau_{jr}^{2,m}}{n_2^m}
\end{align*}
\end{block}
}
\end{frame}
\section{Model selection}
\begin{frame}
\frametitle{Problem of choosing $(Q_1, Q_2)$}
Need to select $Q_1$ and $Q_2$. BIC-Like criterion\footnote{ICL + Entropy + penalty}
\begin{align*}
\text{BIC-L}(\bm{Y}, Q_1, Q_2) & = \max_{\theta} \mathbb{E}_{\mathcal{R}_{\mathbf{Y},\hat{\tau}}} [\ell_c(\bm{Y,Z,W};\theta)] + \mathcal{H(\mathcal{R}_{\mathbf{Y},\hat{\tau}})} - \frac{1}{2}\text{pen}(\theta, Q_1, Q_2) \\
& = \max_{\theta} \mathcal{J(\mathcal{R}_{\mathbf{Y},\hat{\tau}}, \theta)} - \frac{1}{2}\text{pen}(\theta, Q_1, Q_2)
\end{align*}
\begin{alertblock}{Exploration problems}
\begin{itemize}
\item Exploration of $\mathbb{N}^2$ costly.
\item Sensitivity to initializations.
\end{itemize}
\end{alertblock}
\end{frame}
\begin{frame}
\frametitle{Choice of $(Q_1,Q_2)$ - Greedy approach}
\begin{columns}
\begin{column}{0.5\linewidth}
\begin{tikzpicture}
\input{tikz/greedy-exploration.tex}
\end{tikzpicture}
\end{column}
\begin{column}{0.35\linewidth}
\begin{itemize}
\item Initial model~:\\
\begin{tikzpicture}
\draw[fill=gray, draw=gray] circle [radius=0.225cm];
\end{tikzpicture}
\onslide<2->{
\item Model after \emph{split}~:
\begin{tikzpicture}
\draw[fill=blueind, draw=blueind] circle [radius=0.225cm];
\end{tikzpicture}
\item Model maximizing the criterion~:\\
\begin{tikzpicture}
\draw[fill=white, draw=green, very thick] circle [radius=0.225cm];
\end{tikzpicture}
}
\onslide<3->{
\item Model after \emph{merge}~:
\begin{tikzpicture}
\draw[fill=red, draw=red] circle [radius=0.225cm];
\end{tikzpicture}
}
\end{itemize}
\end{column}
\end{columns}
\end{frame}
\begin{frame}
\frametitle{Choice of $(Q_1,Q_2)$ - Sliding window}
\begin{columns}
\begin{column}{0.6\textwidth}
\begin{figure}
\input{tikz/moving-window}
\caption{Sliding window}
\end{figure}
\end{column}
\begin{column}{0.4\textwidth}
\only<3>{\begin{block}{}
Initialization of the model if necessary
\end{block}}
\only<9>{\begin{block}{}
Localization of the new mode
\end{block}}
\only<10>{\begin{block}{}
Move to the new mode then iterate
\end{block}}
\end{column}
\end{columns}
\end{frame}
\section{Application}
\label{sec:application}
\begin{frame}
\frametitle{Results~\cite{baldockSystemsApproachReveals2019}}
\only<1>{
\begin{figure}[ht]
\centering
\begin{subfigure}[t]{0.5\textwidth}
\centering
\includegraphics[width=0.45\textwidth]{tikz/applications/baldock/mat-Baldock2019_Bristol.pdf}
\caption{Bristol}
\end{subfigure}\hfil
\begin{subfigure}[t]{0.5\textwidth}
\centering
\includegraphics[width=0.45\textwidth]{tikz/applications/baldock/mat-Baldock2019_Edinburgh.pdf}
\caption{Edinburgh}
\end{subfigure}
\newline
\begin{subfigure}[ht]{0.5\textwidth}
\centering
\includegraphics[width=0.5\textwidth]{tikz/applications/baldock/mat-Baldock2019_Leeds.pdf}
\caption{Leeds}
\end{subfigure}\hfil
\begin{subfigure}[ht]{0.5\textwidth}
\centering
\includegraphics[width=0.5\textwidth]{tikz/applications/baldock/mat-Baldock2019_Reading.pdf}
\caption{Reading}
\end{subfigure}
\caption{Adjacency matrices,~\cite{baldockSystemsApproachReveals2019}}
\end{figure}
}
\only<2>{
\begin{figure}[ht]
\centering
\begin{subfigure}[t]{0.5\textwidth}
\centering
\includegraphics[width=0.45\textwidth]{tikz/applications/baldock/colbisbm-mat-Baldock2019_Bristol.pdf}
\caption{Bristol}
\end{subfigure}\hfil
\begin{subfigure}[t]{0.5\textwidth}
\centering
\includegraphics[width=0.45\textwidth]{tikz/applications/baldock/colbisbm-mat-Baldock2019_Edinburgh.pdf}
\caption{Edinburgh}
\end{subfigure}
\newline
\begin{subfigure}[ht]{0.5\textwidth}
\centering
\includegraphics[width=0.5\textwidth]{tikz/applications/baldock/colbisbm-mat-Baldock2019_Leeds.pdf}
\caption{Leeds}
\end{subfigure}\hfil
\begin{subfigure}[ht]{0.5\textwidth}
\centering
\includegraphics[width=0.5\textwidth]{tikz/applications/baldock/colbisbm-mat-Baldock2019_Reading.pdf}
\caption{Reading}
\end{subfigure}
\caption{Reordered adjacency matrices by \emph{iid}-colBiSBM,~\cite{baldockSystemsApproachReveals2019}}
\end{figure}
}
\end{frame}
\begin{frame}
\frametitle{Network clustering}
\begin{figure}[ht]
\includegraphics[width=0.45\textwidth]{tikz/applications/baldock/mat-Baldock2011_TB+Baldock2011_JN.pdf}
\caption{Adjacency matrix,~\cite{baldockDailyTemporalStructure2011}}
\end{figure}
\end{frame}
\begin{frame}[allowframebreaks]
\frametitle{Application to~\cite{baldockDailyTemporalStructure2011,
baldockSystemsApproachReveals2019}}
\begin{figure}[t]
\centering
\begin{subfigure}{0.5\textwidth}
\centering
\includegraphics[scale=0.2,angle=-90]{backup-app-iid.png}
\caption{Model $iid$}
\end{subfigure}%
~
\begin{subfigure}{0.5\textwidth}
\centering
\includegraphics[scale=0.2,angle=-90]{backup-app-pirho.png}
\caption{Model $\pi\rho$}
\end{subfigure}%
\caption{Partitioning of networks
of~\cite{baldockDailyTemporalStructure2011,
baldockSystemsApproachReveals2019}}
\end{figure}
\begin{figure}[t]
\centering
\begin{subfigure}{0.5\textwidth}
\centering
\includegraphics[scale=0.1]{backup-app-iid-struct1.png}
\includegraphics[scale=0.2]{backup-app-iid-struct2.png}
\caption{Model $iid$,\\
separate African network and English networks}
\end{subfigure}%
~
\begin{subfigure}{0.5\textwidth}
\centering
\includegraphics[scale=0.2]{backup-app-pirho-struct.png}
\caption{Model $\pi\rho$,\\
merge African and English networks}
\end{subfigure}%
\caption{Structures detected for networks
of~\cite{baldockDailyTemporalStructure2011,
baldockSystemsApproachReveals2019}}
\end{figure}
\end{frame}
\begin{frame}{Clustering algorithm}
\centering
\vspace{0.25\baselineskip}
\begin{tikzpicture}[scale=0.85]
\input{tikz/clustering.tex}
\end{tikzpicture}
\[
D_{\mathcal{M}}(m,m') = \sum_{q = 1}^{Q_1} \sum_{r = 1}^{Q_2} \max(\widetilde{\pi}_{q}^{m}, \widetilde{\pi}_{q}^{m'}) \left( \widetilde{\alpha}_{qr}^{m} - \widetilde{\alpha}_{qr}^{m'}\right)^{2} \max(\widetilde{\rho}_{r}^{m}, \widetilde{\rho}_{r}^{m'})
\]
\end{frame}
\begin{frame}{Results}
\begin{figure}[ht]
\centering
\begin{subfigure}{0.5\textwidth}
\centering
\includegraphics[width=1\textwidth]{tikz/applications/baldock/bisbm-mat-Baldock2011_TB+Baldock2011_JN.pdf}
\caption{Reordered by LBM}
\end{subfigure}\hfil
\begin{subfigure}{0.5\textwidth}
\centering
\includegraphics[width=1\textwidth]{tikz/applications/baldock/pirho-colbisbm-mat-Baldock2011_TB+Baldock2011_JN.pdf}
\caption{Reordered by $\pi\rho$-colBiSBM}
\end{subfigure}
\caption{Reordered adjacency matrix by $\pi\rho$-colBiSBM,~\cite{baldockDailyTemporalStructure2011}}
\end{figure}
\end{frame}
\section{Conclusion}
\label{sec:conclusion}
\begin{frame}
\frametitle{Conclusion and perspectives}
% DONE Add a conclusion perspective slide
% Recall models with clustering
% Mention analysis of corrected networks for sampling
% Link to the package
\begin{block}{Capabilities}
\begin{itemize}
\item 4 models including 3 with flexibility on at least one of
the dimensions (adaptability to data).
\item Detect classic and less classic structures in an agnostic way.
\item Partition a set of networks according to their structures.
\end{itemize}
\end{block}
\end{frame}
\begin{frame}{Perspectives}
\begin{itemize}
\item Investigate stability against randomness and local \emph{optima}.
\item Proof of identifiability of the $\pi\rho$ model.
\end{itemize}
\begin{block}{Package and applications}
\begin{itemize}
\item Integration into the \texttt{colSBM} package, improvement of user interface and
addition of ecologists' feedback
\item CRAN publication
\item Integrate the possibility of an additional criterion for clustering
\item Apply clustering to data from
\cite{pichonTellingMutualisticAntagonistic2024,doreRelativeEffectsAnthropogenic2021}
\end{itemize}
\end{block}
\bigskip
\centering
Thank you for your attention~!
\end{frame}