Research map / Video / Video generation
Grounding World Simulation Models in a Real-World Metropolis
Junyoung Seo, Hyunwook Choi, Minkyung Kwon, Jinhyeok Choi, Siyoon Jin, Gayoung Lee, Junho Kim, JoungBin Lee, Geonmo Gu, Dongyoon Han, Sangdoo Yun, Seungryong Kim and 1 more
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 cross-temporal pairing of references and targets (ablation) | SWM (TF) · Busan-City-Bench | FID: 44.74 -> 28.43 | Improved |
| add synthetic CARLA urban data (ablation) | SWM (TF) · Busan-City-Bench | mPSNR / FID: FID 27.74 -> 28.43; mPSNR 13.52 -> 14.56 | Mixed |
| add geometric referencing (ablation) | SWM (TF) · Busan-City-Bench | FID: 33.01 -> 28.43 | Improved |
| add semantic referencing (ablation) | SWM (TF) · Busan-City-Bench | FID: 30.27 -> 28.43 | Improved |
| add Virtual Lookahead Sink vs no sink (ablation) | SWM (TF) · Busan-City-Bench | FID: 33.06 -> 28.43 | Improved |
| use VL Sink instead of first-frame sink (ablation) | SWM (TF) · Busan-City-Bench | FID: 32.71 -> 28.43 | Improved |
| use VL Sink instead of first-position sink on long horizon (ablation) | SWM · Busan-City-Bench (1460 frames) | FID: 28.57 -> 25.13 | Improved |
| add real street-view data (ablation) | SWM (TF) · Busan-City-Bench | FVD: 467.58 -> 301.76 | Improved |
| use intermittent freeze-frame instead of channel concatenation for view interpolation (ablation) | view interpolation · Waymo test | LPIPS: 0.245 -> 0.162 | Improved |
| reduce retrieved references K from 5 to 1 (ablation) | SWM (TF) · Busan-City-Bench | mPSNR / FID | Mixed |
| use TF vs SF variant of SWM | SWM · Busan vs Ann Arbor | FID: Busan 28.43 (TF) vs 32.50 (SF); Ann Arbor 56.61 (TF) vs 43.97 (SF) | Mixed |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- Yume1.5 — Yume1.5: A Text-Controlled Interactive World Generation Model (CVPR 2026) · better on 1 of 1 shared result
- Aether — Aether: Geometric-Aware Unified World Modeling (ICCV 2025) · better on 1 of 1 shared result
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
- SAMPO: Scale-wise Autoregression with Motion Prompt for Generative World Models (NeurIPS 2025 · similarity 0.86)
- From Zero to Hero: Training-Free Custom Concept Spawning in World Models (NeurIPS 2026 · similarity 0.86)
- Sekai: A Video Dataset towards World Exploration (NeurIPS 2025 · similarity 0.86)
- SceneDiffuser++: City-Scale Traffic Simulation via a Generative World Model (CVPR 2025 · similarity 0.86)
- stable-worldmodel: A Platform for Reproducible World Modeling Research and Evaluation (NeurIPS 2026 · similarity 0.85)
- WorldModelBench: Judging Video Generation Models As 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.
