A 2-day internal brainstorming workshop of the SHARP project was held at ENS de Lyon on January 24-25 2025. Each day started with an invited seminar open to all registered participants
- Anne-Laure Ligozat “How to evaluate the carbon footprint in AI projects”
- Guillaume Charpiat “Neural Network Growth for Frugal AI”
The rest of the workshop was articulated around spotlight presentations, plenary discussions to discuss project-wide research topics, and brainstorming sessions in smaller working groups. The list of spotlights was :
- Multi-Level Information Retrieval Augmented Generation for Knowledge-based Visual Question Answering (EMNLP 2024) [Adjali, Ferret, Ghannay, Le Borgne]
- FOTON: Forward Only Training of Orthogonal Networks [Alex Colagrande]
- A path-norm toolkit for modern networks: consequences, promises and challenges. ICLR 2024 [Gonon, Brisebarre, Riccietti, Gribonval]
