< Back to events




18 November 2024
Geometry-Grounded Representation Learning
Title: Geometry-Grounded Representation Learning
Speaker: Erik Bekkers
When: Monday 18 November 15-16:00
Where: LAB42 L3.36, Online: https://uva-live.zoom.us/j/81510587923?pwd=bXbUElREqUydae5QQ3FudeRrakbO86.1
This talk introduces new ideas for learning reliable representations by explicitly preserving geometric structure. I'll present recent work in which we treat data as fields with geometric latent variables, modeled through equivariant neural fields [1]. The equivariance relation ensures that the latent variables are geometrically meaningful, which we show through various classification and segmentation experiments. I subsequently show that this allows for continuous PDE forecasting entirely in latent space [2]. That is, instead of modeling a PDE on the fields, we model equivariant dynamics through equivariant neural ODEs applied to the geometric latent variables. I'll further highlight a generalized notion of weight-sharing as sharing operations over equivalence classes of point-pairs [3]. Through this principle, we developed a general-purpose equivariant neural network that achieves state-of-the-art performance on a wide range of tasks, from regression to segmentation and the generation of novel molecules. I conclude with our most recent work on designable (and thus interpretable) latent space through isometry learning and pull-back geometry [4].
[1] Wessels, D. R., Knigge, D. M., Papa, S., Valperga, R., Vadgama, S., Gavves, E., & Bekkers, E. J. (2024). Grounding Continuous Representations in Geometry: Equivariant Neural Fields. arXiv preprint arXiv:2406.05753. ISO 690
[2] [accepted at Neurips 2024] Knigge, D. M., Wessels, D. R., Valperga, R., Papa, S., Sonke, J. J., Gavves, E., & Bekkers, E. J. (2024). Space-Time Continuous PDE Forecasting using Equivariant Neural Fields. arXiv preprint arXiv:2406.06660.
[3] Bekkers, E. J., Vadgama, S., Hesselink, R., Van der Linden, P. A., & Romero, D. W. Fast, Expressive $\mathrm {SE}(n) $ Equivariant Networks through Weight-Sharing in Position-Orientation Space. In The Twelfth International Conference on Learning Representations. ISO 690
[4] de Kruiff, F., Bekkers, E., Öktem, O., Schönlieb, C. B., & Diepeveen, W. (2024). Pullback Flow Matching on Data Manifolds. arXiv preprint arXiv:2410.04543
Aankomend >
Upcoming events

4 April 2025
ELSA Event 2025: Exploring AI’s Ethical, Legal & Societal Impact
Amsterdam AI invites you to the 2025 ELSA event, bringing together a diverse group of stakeholders to explore the ethical, legal, and societal aspects of artificial intelligence. This exclusive, invite-only event offers a unique opportunity to connect, collaborate, and gain insights into the latest ELSA research and funding opportunities.
read more >

April 10
Masterclass AI Foundation Models
Discover the latest advancements in AI foundation models during this one-day masterclass on April 10, 2025, at Amsterdam Science Park. Led by AI expert Prof. Dr. Cees Snoek, this session will provide insights into the capabilities and societal impact of these groundbreaking AI models.
read more >

15 & 16 April
ICT. OPEN 2025 – Call for Abstracts!
Share your latest research with leading minds in the field of ICT? The annual NWO ICT.OPEN2025 conference is inviting researchers to submit abstracts for demo, oral, and poster presentations. This is a opportunity to showcase your work to peers, industry leaders, and academics from across the Netherlands.
read more >