GeUmetric Deep Learning Workshop 2025

Day 1, Tuesday 19/8

13:00-13:30 Welcome

13:30-14:00 Oscar Carlsson
Differential Geometry in Equivariant CNNs via Biprincipal Bundles

14:00-14:30 Elias Nyholm
Unifying transformers and CNNs as equivariant maps

14:30-15:00 Emma Andersdotter Svensson
A Bundle Formalism for Equivariant Neural ODEs

15:00-15:30 Coffee break

15:30-16:00 Max Guille
Finite-width Neural Tangent Kernels from Feynman Diagrams

16:00-16:30 Philipp Misof
Equivariant Neural Tangent Kernels - Connecting data augmentation and equivariant architectures

16:30-17:00 Oskar Nordenfors
Data augmentation yields equivariance for ensembles of neural networks

Day 2, Wednesday 20/8

09:00-10:00 Elisenda Grigsby
Local complexity measures in modern parameterized function classes for supervised learning

10:00-10:30 Coffee break

10:30-11:00 Vahid Shahverdi
Neuroalgebraic Cartography: Geometry Under the Surface

11:00-11.30 Stefano Mereta
The geometry of the neuromanifold of ReLU networks

11:30-12:00 Alex Massarenti
On the dimension of neuro varieties

12:00-13:30 Lunch

13:30-14:00 Felix Stollenwerk
Research at AI Sweden

14:00-14:30 Hampus Linander
PEAR: Equal Area Weather Forecasting on the Sphere

14:30-15:00 Tino Paulsen
Hyperbolic Music Representations

15:00-15:30 Coffee break

15:30-16:00 WASP Cluster meeting

16:30- Social event + dinner

Day 3, Thursday 21/8

09:00-10:00 Luca Cosmo
Graph Generative Models for Interpretable Graph Neural Networks

10:00-10:30 Coffee break

10:30-11:00 Longde Huang
Learning Chern Numbers of Topological Insulators With Gauge Equivariant Neural Networks

11:00-11.30 Alexander Friedrich
Autoencoding via Neural ODEs on M-polyfolds

11:30-12:00 Leonard Mushunje
TBA

Book of Abstracts

GeUmetric Deep Learning Workshop 2025