Differentiable Hierarchical Visual Tokenization

NeurIPSSpotlight2025

Authors
Marius Aasan, Martine Hjelkrem-Tan, Nico Catalano, Changkyu Choi, Adín Ramírez Rivera
Venue
NeurIPS 2025
Track
Spotlight

TL;DR

Vision Transformers rely on fixed patch tokens that ignore the spatial and semantic structure of images…

Opening excerpt from the authors’ abstract. source

Read the paper

Topics

transformer

Related NeurIPS papers

← All NeurIPS 2025 Spotlight papers · Browse the whole archive