TencentARC/GAE-GeometricAutoEncoder
[arxiv'26] GAE: Learning a Geometry-Native Latent Space for 3D-Consistent World Generation
📚 Overview
GAE (Geometric … ) is a research‑level deep‑learning model that learns a geometry‑native latent space …
...
Related
- Project
lukasHoel/video_to_worldA 3D reconstruction pipeline that resolves geometric inconsistencies in videos generated by video diffusion models to create stable 3D worlds.
- Project
microsoft/LatentSpatialMemoryA framework for video world models that stores 3D scene content as latent tokens to improve generation speed and reduce memory overhead by avoiding repeated RGB rendering.
- Project
haoz19/world-tracingWorld Tracing is a generative AI project that predicts multilayer 3D point clouds from images or video, capturing both visible and occluded geometry.
- Project
nicolasvonluetzow/GaussianGPTGaussianGPT is a transformer-based model that generates 3D Gaussian scenes via next-token prediction, supporting controllable 3D scene generation, completion, and outpainting.
- Dispatch
Genie 3 real-time world model releaseGenie 3 is DeepMind’s new real‑time world model that generates interactive 720p environments at 24 fps and maintains consistency for minutes, enabling advanced embodied‑agent research.