Paper Club #2 - Oriane Siméoni
Join us on Thursday, January 29th for the second edition of the Unaite Paper Club, with Oriane Siméoni working on DINO at Meta
Self-supervised learning has directed much of the recent progress in computer vision, with impressive results on a broad array of tasks. Instead of relying on labeled datasets, these methods learn meaningful representations directly from raw visual data, without additional annotation. Among them, DINO models developed by Meta are currently one of the most universal approaches to efficiently represent images and videos.
For this edition of the Paper Club, Unaite is pleased to host Oriane Siméoni, from the DINO team at Meta. This seminar will discuss the core ideas behind the DINOv3 model with key insights, technical takeaways, and open research directions in self-supervised visual representation learning.
The session will be followed by a cocktail !