4 papers
Self-Supervised Representation Learning via Hyperspherical Density Shaping
Esteban RodrÃguez-Betancourt, Edgar Casasola-Murillo
Modern self-supervised representation learning methods often relies on empirical heuristics that are not theoretically grounded. In this study we propose HyDeS, a theoretically gro…
Geometric Analysis of Self-Supervised Vision Representations for Semantic Image Retrieval
Esteban RodrÃguez-Betancourt, Edgar Casasola-Murillo
Content-based image retrieval (CBIR) systems enable users to search images based on visual content instead of relying on metadata. The text domain has benefited from vector search…
Randomly Initialized Networks Can Learn from Peer-to-Peer Consensus
Esteban RodrÃguez-Betancourt, Edgar Casasola-Murillo
In self-supervised learning, self-distilled methods have shown impressive performance, learning representations useful for downstream tasks and even displaying emergent properties.…
Hypersolid: Emergent Vision Representations via Short-Range Repulsion
Esteban RodrÃguez-Betancourt, Edgar Casasola-Murillo
A recurring challenge in self-supervised learning is preventing representation collapse. Existing solutions typically rely on global regularization, such as maximizing distances, d…