GAE: A Geometry-Native Latent Space for 3D-Consistent Generation

Tencent ARC and CUHK researchers introduced GAE, a geometry-native autoencoder that compresses geometric foundation features into a shared latent space where flow models directly generate video and 3D point clouds, cutting FVD by over 20% and halving camera error.

2026-09-22 ~ 2026-09-23 · 4 related posts

• GAE: A Geometry-Native Autoencoder Cuts World Model FVD by 23.1% — yshan2u · 2026-09-22
• GAE generates video and 3D geometry natively in a shared geometry latent space, code released — yshan2u · 2026-09-22
• Tencent ARC's GAE: geometry-native latent space cuts FVD by up to 23% in world generation — TencentARC · 2026-09-23
• GAE Paper: Geometry-Native Latent Space Cuts FVD by 23% and Halves Camera Error — yshan2u · 2026-09-23