Overview

The LoG Meetup in Lausanne is associated with the virtual Learning on Graphs (LoG) Conference, an annual research conference with focus on graph machine learning and geometric deep learning.  To complement the main event with an offline social venue, we bring together computer scientists from the lake Geneva region and all over Switzerland to discuss machine learning on graphs and geometry, from the theory to its application across disciplines.

The meetup will feature five invited talks, followed by a poster session and apéro, providing plenty of opportunities to discuss all things graphs and geometry. To showcase research efforts from the local community, boost interdisciplinary research and foster collaborations, there will be also a poster session. Whether you are an expert or just curious to hear how graphs could be applied in your area, everyone is welcome!

Key Information

• What: Free to attend and open to everyone Learning on Graphs Meetup

• When: 22nd of November, 2023, 2pm - 7pm

• Where: EPFL, Lausanne, BM 5202 (5th floor)

• Format: Offline

Registration

If you plan to attend the meetup, please register here: https://forms.gle/c4HjeuDoBTpXN95F9

Update (16/11): Due to a large enthusiasm, the event has now reached maximal capacity, and registrations are closed until further notice!

Program

• 14:00 - 14:15: Introduction

• 14:15 - 14:55: Charlotte Bunne. "Neural Optimal Transport and Flow Matching for Single-Cell Biology"

• 14:55 - 15:35: Clément Vignac. "Discrete denoising diffusion models for graph generation"

• 15:35 - 15:55: Mehmet Akdel. "DiffMaSIF: score-based diffusion models for protein surfaces"

• 15:55 - 16:20: Break

• 16:20 - 17:00: Dorina Thanou. "Graph representations for biology and medicine: Challenges and applications"

• 17:00 - 17:40: Michaël Defferrard. "Leveraging topology, geometry, and symmetries for efficient Machine Learning"

• 17:40 - 19:30: Poster session

Posters

• Vignesh Ram Somnath: “DockGame: Cooperative Games for Multimeric Rigid Protein Docking”

• Kenneth Atz: “Deep interactome learning for de novo drug design”

• Christian Koke: “HoloNets: Spectral Convolutions do extend to Directed Graphs”

• Davide Ghio: “Bayes-optimal inference for spreading processes on random networks”

• Ksenia Briling: “EquiReact: an equivariant neural network for chemical reactions”

• Elizaveta Kozlova: “Inpainting Protein Sequence and Structure with ProtFill”

• Lucien Krapp: “A Geometric Transformer for Structural Biology: from binding interface prediction to protein design”

• Andrei Catalin Coman: “Transformers as Graph-to-Graph Models”

• Odilon Duranthon: “Generalization in Graph Convolution Networks”

• Oisin Morrison: “Learning operators through graph neural networks”

• Juan F. Flórez-Ospina: “Rank-order path graphs and its use in compressive spectral imaging with side information”

• Maria Boulougouri: “Molecular Set Representation Learning”

• Yoann Boget: “Discrete Graph Auto-encoder”

• Gregor Krzmanc: “Towards Particle Flow Event Reconstruction at the Future Circular Collider with GNNs”

• Mengjie Zhao: “Dynamic Edge via Graph Attention for Early Fault and Detection in Complex Systems”

• Sacha Raffaud: “TS-DiffuGen & TS-Flow: Generative Models for Reaction Transition State conformation generation”

• Florian Grötschla: “SALSA-CLRS: A Sparse and Scalable Benchmark for Algorithmic Reasoning”

Organizers