Research map / Video / Video generation
LIVE: Long-horizon Interactive Video World Modeling
Junchao Huang, Ziyang Ye, Xinting Hu, Tianyu He, Guiyu Zhang, Shaoshuai Shi, Jiang Bian, Li Jiang
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 cycle-consistency (reverse generation) objective (ablation) | LIVE · RealEstate10K 0-64 frames | PSNR: 13.99 -> 18.11 | Improved |
| add cycle-consistency (reverse generation) objective (ablation) | LIVE · RealEstate10K 0-200 frames | PSNR: 11.18 -> 14.57 | Improved |
| add random per-frame timestep noise on rollout context vs none (ablation) | LIVE · RealEstate10K 0-200 frames | PSNR: 13.83 -> 14.57 | Improved |
| add random per-frame timestep noise vs fixed-scale noise (ablation) | LIVE · RealEstate10K 0-200 frames | PSNR: 14.09 -> 14.57 | Improved |
| add random timestep noise vs none (ablation) | LIVE · RealEstate10K 0-64 frames | PSNR: 17.76 -> 18.11 | Improved |
| use progressive curriculum decreasing p instead of fixed p=1 (ablation) | LIVE · RealEstate10K 0-64 frames | PSNR: 16.78 -> 18.11 | Improved |
| use progressive curriculum instead of fixed p=1 (ablation) | LIVE · RealEstate10K 0-200 frames | LPIPS: 0.4800 -> 0.4163 | Improved |
| use LIVE training instead of NFD-DF | NFD architecture · RealEstate10K >=256 frames | PSNR: 11.51 -> 13.89 | Improved |
| use LIVE training instead of NFD-DF | NFD architecture · Minecraft 0-200 frames | PSNR: 12.34 -> 14.02 | Improved |
| use LIVE training instead of NFD-DF | NFD architecture · UE Engine >=400 frames | PSNR: 11.02 -> 12.96 | Improved |
| use DF instead of TF | NFD architecture · RealEstate10K 0-200 frames | PSNR: 11.63 -> 12.21 | Improved |
Closest papers by meaning
- WorldWeaver: Generating Long-Horizon Video Worlds via Rich Perception (NeurIPS 2025 · similarity 0.90)
- Learning World Models for Interactive Video Generation (NeurIPS 2025 · similarity 0.90)
- WorldReasonBench: Human-Aligned Stress Testing of Video Generators as Future World-State Predictors (NeurIPS 2026 · similarity 0.89)
- Video World Models with Long-term Spatial Memory (NeurIPS 2025 · similarity 0.88)
- LongScape: Advancing Long-Horizon Embodied World Models with Context-Aware MoE (NeurIPS 2026 · similarity 0.88)
- SlowFast-VGen: Slow-Fast Learning for Action-Driven Long Video Generation (ICLR 2025 · similarity 0.88)
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.
