When? 18 June, 14:30 - 18:00
Where? Québec City Convention Center (Room 206A)
We are proud to present a high-quality program with speakers from different communities. Speakers who want to publish their talks after the satellite can send their slides via E-Mail.
| JUNE 18th | Presenter |
|---|---|
| 14:30 - 14:45 | Ingo Scholtes Opening Statement |
| 14:45 - 15:30 |
Tina Eliassi-Rad
Northeastern University, USA Title: Hypergraph Mining: Patterns, Tools, and Generators Abstract TBA |
| 15:30 - 16:00 |
Philipp Hövel
Saarland University, Germany Title: Information Parity and Motif Analysis to Fingerprint Networks and their Symmetries Abstract TBA |
| 16:00 - 16:30 | Coffee Break |
| 16:30 - 17:00 |
Veronica Lachi
Fondazione Bruno Kessler, Italy Title: Graph Neural Networks for Temporal Graphs: State of the Art, Open Challenges, and Opportunities Abstract TBA |
| 17:00 - 17:30 |
Bruno Ribeiro
Purdue University, USA Title: Enhancing AI Robustness Through Hypergraph Learning In this talk we explore the use of hypergraphs to create more robust machine learning models. We will start with how hypergraphs help design neural networks that can better reason algorithmically. Then, we will see how higher-order networks are used today for state-of-the-art robustness to out-of-distribution shifts in graph tasks. Finally, we introduce how to learn hypergraph embeddings from static dyadic graphs for better multi-link predictions tasks. |
| 17:30 - 18:00 |
Vincent P. Grande
RWTH Aachen University, Germany Title: Hodge Learning: What the Eigenspectrum of the Hodge Laplacian tells us about higher-order network topology and geometry The rich spectral information of the graph Laplacian has been instrumental in graph theory, machine learning, and graph signal processing for a diverse range of applications. Recently, Hodge Laplacians have come into focus as a generalisation of the Laplacian for simplicial complexes. Many authors analyse the smallest eigenvalues of the Hodge Laplacian, which are connected to important topological and spatial properties of simplicial complexes and higher-order networks. However, small eigenvalues of the Hodge Laplacians can carry different information depending on whether they are related to curl or gradient eigenmodes. We track individual harmonic, curl, and gradient eigenvectors/-values through the so-called persistence filtration, leveraging the full information contained in the spectrum across all scales of a point cloud. Finally, we use our insights (a) to introduce Hodge spectral clustering and (b) to classify edges and higher-order simplices based on their relationship to the smallest harmonic, curl, and gradient eigenvectors. |
| 18:00 | Organisers Closing Statement |