Research map / Video / Video generation
Aether: Geometric-Aware Unified World Modeling
Haoyi Zhu, Yifan Wang, Jianjun Zhou, Wenzheng Chang, Yang Zhou, Zizun Li, Junyi Chen, Chunhua Shen, Jiangmiao Pang, Tong He
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add video depth reconstruction objective to multi-task training (ablation) | Aether · action-conditioned navigation | PSNR: 18.97 -> 19.70 | Improved |
| add video depth reconstruction objective (ablation) | Aether · action-conditioned navigation | LPIPS: 0.3074 -> 0.2659 | Improved |
| add video depth reconstruction objective (ablation) | Aether · action-free visual path planning | VBench weighted average: 79.59 -> 80.67 | Improved |
| post-train CogVideoX with geometric multi-task data | Aether · video prediction without action | VBench weighted average: 78.51 -> 80.04 | Improved |
| add camera trajectory raymap action conditioning with post-training | Aether · action-conditioned video prediction | VBench weighted average: 79.92 -> 80.71 | Improved |
| use Aether instead of CUT3R for KITTI depth | Aether · KITTI video depth | Abs Rel: 0.118 -> 0.056 | Improved |
| use Aether instead of MonST3R-GA on BONN depth | Aether · BONN video depth | Abs Rel: 0.067 -> 0.273 | Hurt |
| use Aether instead of CUT3R for camera pose on ScanNet | Aether · ScanNet camera pose | ATE: 0.099 -> 0.176 | Hurt |
| use Aether instead of CUT3R for camera pose on Sintel | Aether · Sintel camera pose | ATE: 0.213 -> 0.189 | Improved |
| use Aether instead of DepthCrafter for depth on BONN (scale&shift) | Aether · BONN video depth | Abs Rel: 0.253 -> 0.308 | Hurt |
Later papers that compared with this method
6 later accepted papers report Aether as a baseline in their results tables; 6 of them report a better number on most of the results they share with it.
- Stereo World Model: Camera-Guided Stereo Video Generation (CVPR 2026 · better on 4 of 4 shared results)
- How Much 3D Do Video Foundation Models Encode? (CVPR 2026 · better on 4 of 4 shared results)
- OctWorld: Long-Range World-Consistent Video Generation with Octree-based 3D Mapping (ECCV 2026 · better on 3 of 3 shared results)
- VGGT-World: Transforming VGGT into an Autoregressive Geometry World Model (ECCV 2026 · better on 2 of 2 shared results)
- UnityVideo: Unified Multi-Modal Multi-Task Learning for Enhancing World-Aware Video Generation (CVPR 2026 · better on 1 of 1 shared result)
- Grounding World Simulation Models in a Real-World Metropolis (ECCV 2026 · better on 1 of 1 shared result)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- DepthCrafter — DepthCrafter: Generating Consistent Long Depth Sequences for Open-world Videos (CVPR 2025) · better on 2 of 3 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- DriveDreamer-Policy: A Geometry-Grounded World-Action Model for Unified Generation and Planning (NeurIPS 2026 · similarity 0.87)
- EA-WM: Event-Aware Generative World Model with Structured Kinematic-to-Visual Action Fields (NeurIPS 2026 · similarity 0.86)
- MindJourney: Test-Time Scaling with World Models for Spatial Reasoning (NeurIPS 2025 · similarity 0.85)
- UniVR: Thinking in Visual Space for Unified Visual Reasoning (NeurIPS 2026 · similarity 0.85)
- SAMPO: Scale-wise Autoregression with Motion Prompt for Generative World Models (NeurIPS 2025 · similarity 0.85)
- Learning 3D Persistent Embodied World Models (NeurIPS 2025 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
