New Method Learns Directional Randers Geometry for Latent-Space Exploration
A new arXiv paper proposes learning directional Randers geometry to improve exploration of latent spaces learned by deep neural networks from images, addressing limitations in downstream analysis.
Read the full story at arXiv cs.LG ↗
Timeline · 1 report
- 2026-11-04 05:00 · arXiv cs.LG
Learning Directional Randers Geometry for Latent-Space Exploration