Louis/Thèse/Lectures/@neumannBipartiteStochasticBlock2018.md
2026-06-09 14:45:18 +02:00

91 lines
2.3 KiB
Markdown
Raw Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

---
categories: [literature note, ]
title: Note de lecture de *Bipartite Stochastic Block Models with Tiny Clusters* de
Stefan Neumann.
bibliography: ../these_ref.bib
---
::: {.callout-note title="Synthèse"}
**Contribution**::
**Related**::
:::
::: {.callout-note title="Markdown"}
**FirstAuthor**:: Neumann, Stefan
**Title**:: Bipartite Stochastic Block Models with Tiny Clusters
**Year**:: 2018
**Citekey**:: neumannBipartiteStochasticBlock2018
**itemType**:: conferencePaper
**Volume**:: 31
**Publisher**:: Curran Associates, Inc.
:::
::: {.callout-note title="Pièces-jointes"}
- [Full Text PDF](file:///home/louis/snap/zotero-snap/common/Zotero/storage/NNDNV2GV/Neumann%20-%202018%20-%20Bipartite%20Stochastic%20Block%20Models%20with%20Tiny%20Clusters.pdf).
:::
::: {.callout-note title="Abstract"}
::::
# Prise de notes
{{< include local_macros.tex.md >}}
![[local_macros.tex]]
%% begin user_notes %%
Vaguement en lien avec notre sujet de papier.
Propose une méthode pour détecter des petits clusters du "côté droit" du graphe (les noeuds $j \in V$). Avec des tailles de clusters de l'ordre de $n_{2}^{\varepsilon}$ où $\varepsilon>0$
La preuve de la proposition 4 sur la récupération des clusters de $V$ est intéressante.
%% end user_notes %%
# Annotations importées
%% begin annotations %%
## Importées : 2026-05-22 3:23 pm
<mark style="background-color: #ffd400">Quote</mark>
> For V there are clusters V1, … , Vk with Vi ⊆ V ; we do not assume that the Vj are disjoint or that their union equals V
<mark style="background-color: #ff6666">Quote</mark>
> Fix two probabilities p > q ≥ 0. For any two vertices u ∈ Ui and v ∈ Vi, insert an edge with probability p; for u ∈ Ui and v 6∈ Vi, insert an edge with probability q
## Importées : 2026-05-22 3:31 pm
<mark style="background-color: #5fb236">Quote</mark>
> We make the decision for the parameter based on the likelihood of observing Zv edges incident upon v. Parameter p is more likely if: |Ui
1
1 q |Ui|
Ui
Zv pZv (1 p)|Ui|Z
q Zv 1
1 q
|U
qZv (1 q)|Ui|
v Solving this inequality for Zv gives that one should decide for parameter p if Zv ≥ θ|Ui|, where θ
as in Equation
%% end annotations %%
%% Import Date: 2026-05-22T15:31:44.079+02:00 %%