List of talks
Alexander Chervov : Machine learning methods for Cayley graphs pathfiding
Alexei Kotov : Generalized signature method with applications
Diaaeldin Taha : Demystifying Topological Message-Passing with Relational Structures: A Case Study on Oversquashing in Simplicial Message-Passing
Edward Pearce-Crump : Permutation Equivariant Neural Networks for Symmetric Tensors
Elias Tsigaridas : Randomized geometric tools for anomaly detection in stock markets
Elodie Laine : PETITMOT: A Novel Framework for Inferring Protein Motions from Sparse Data Using SE(3)-Equivariant Graph Neural Networks
Grégoire Sergeant-Perthuis : Inference of Factor Graphs under Topological Transformations
Haochen Yang : Topological shape descriptors with isometry invariance
Jiayi Li : Geometry of Neural Networks with Algebraic Activations
Joseph Nardin-Gennequin : Geometry of protein structures for classification and phylogeny
Jules Tsukahara : Bayesian Information Criterion of singular bi-parametric models over the probability simplex
Matias Bender : Computing minimal presentations of multi-parameter persistent homology
Olga Anosova : A complete invariant of protein backbones exposes thousands of duplicate chains in the PDB.
Renata Turkes : Shoving tubes through shapes gives a sufficient and efficient shape statistic
Sergei Grudinin : Equivariant Fourier-space architectures for volumetric data processing
Serguei Barannikov : Bar-codes, Comparative Bar-codes and Their Applications in Machine Learning.
Susovan Pal : Optimal lifts and their roles in the asymptotics of inference on shapes
Tolga Birdal : Topological Deep Learning
Umut Simsekli : Topological Generalization Bounds for Discrete-Time Stochastic Optimization Algorithms
Vadim Lebovici : A unified theory of integral transforms for shape analysis
Vitaliy Kurlin : Geometry on moduli spaces of real data objects
Yiannis Vlassopoulos : Large Language Models, Tropical Geometry and Directed Metric Spaces
Schedule of the conference: talks are 35 minutes long + 10 minutes of questions
There is a whiteboard but no blackboard