5 papers
Unbiased Open World Regularization for Fair Self-Supervised Learning
L{é}o Nicollier, Marc Pic, Pablo Mus{é} +2
Despite recent advances, self-supervised learning (SSL) models and Joint-Embedding Predictive Architectures (JEPAs) remain susceptible to learning spurious biases in the dataset. T…
Expanding SPHERE-JEPA: A Family of Statistical Regularizers for the Hypersphere
Léo Nicollier, Enric Meinhardt-Llopis, Max Dunitz +3
In Self-Supervised Learning (SSL), preventing representation collapse by explicitly enforcing a uniform distribution on the unit hypersphere has proven to be effective. However, cu…
SPHERE-JEPA: Spherical Prediction with Homogeneous Embeddings
Léo Nicollier, Max Dunitz, Marc Pic +3
A fundamental open question in self-supervised learning (SSL) is the explicit characterization of the optimal geometry of the learned representations. Recently, LeJEPA identified i…
Unlocking Few-Shot Capabilities in LVLMs via Prompt Conditioning and Head Selection
Adhemar de Senneville, Xavier Bou, Jérémy Anger +2
Current Large Vision Language Models (LVLMs) excel at many zero-shot tasks like image captioning, visual question answering and OCR. However, these same models suffer from poor per…
Comparative analysis of dual-form networks for live land monitoring using multi-modal satellite image time series
Iris Dumeur, Jérémy Anger, Gabriele Facciolo
Multi-modal Satellite Image Time Series (SITS) analysis faces significant computational challenges for live land monitoring applications. While Transformer architectures excel at c…