Uploaded April 2026 | Updated September 2026, 3 weeks ago
This talk provides a brief introduction to statistical network analysis and random graph models. We then focus on the problem of estimating the graphon function, which characterizes nonparametric exchangeable random graph models. Our main emphasis is on the setting where multiple networks are observed, which introduces additional challenges compared to the classical single-network framework. To address this, we propose a new histogram-based estimator with low computational complexity. The key idea is to jointly align the nodes across all observed graphs, rather than processing each network independently as in most existing approaches. We establish consistency results for the proposed estimator and demonstrate through numerical experiments that it outperforms current methods in both estimation accuracy and computational efficiency. Finally, we show that, when used for data augmentation in graph neural network classification tasks, our approach leads to improved performance on various real-world datasets. This is joint work with Roland Sogan.
Tabea Rebafka (AgroParisTech)
===
Find this and many more scientific videos on carmin.tv - a French video platform for mathematics and their interactions with other sciences offering extra functionalities tailored to meet the needs of the research community.
===
This talk provides a brief introduction to statistical network analysis and random graph models. We then focus on the problem of estimating the graphon function, which characterizes nonparametric exchangeable random graph models. Our main emphasis is on the setting where multiple networks are observed, which introduces additional challenges compared to the classical single-network framework. To address this, we propose a new histogram-based estimator with low computational complexity. The key idea is to jointly align the nodes across all observed graphs, rather than processing each network independently as in most existing approaches. We establish consistency results for the proposed estimator and demonstrate through numerical experiments that it outperforms current methods in both estimation accuracy and computational efficiency. Finally, we show that, when used for data augmentation in graph neural network classification tasks, our approach leads to improved performance on various real-world datasets. This is joint work with Roland Sogan.
Tabea Rebafka (AgroParisTech)
===
Find this and many more scientific videos on carmin.tv - a French video platform for mathematics and their interactions with other sciences offering extra functionalities tailored to meet the needs of the research community.
===
