A project meeting was held at ISIR (Paris) on July 4th 2024. The day was organized around several spotlight presentations as well as a tutorial.
- Morning spotlights
- “Learning to Simplify Large Computer Vision Problems” [Loic Landrieu]
- “Segmenting Historical Maps with Modern Labels” [Loic Landrieu]
- “Understanding diffusion models” [David Picard]
- “Efficient kernel vector products for optimal transport” [Titouan Vayer]
- “Gradient Aligned, Projected Error (GRAPE)” [Paul Caillon]
- “Chacal : Chain and Causal attention for efficient entity tracking”, [Erwan Fagnou]
- Introduction / tutorial on pyfaust (matrix compression via butterfly factorization) [Pascal Carrivain]
[notebook de présentation de Fast-Deformable-Butterfly (fdb()dans pyfaust)] - Afternoon spotlights
- “Operator Learning with Neural Fields: Tackling PDEs on General Geometries” [Louis Serrano]
- “Physics-Informed Model and Hybrid Planning for Efficient Dyna-Style Reinforcement Learning” [Zakariae El Asri]
- “What Makes Multimodal In-Context Learning Work?” [Folco Bertini]
