Research map / Video / Video generation
Yume1.5: A Text-Controlled Interactive World Generation Model
Xiaofeng Mao, Zhen Li, Chuanhao Li, Xiaojie Xu, Kaining Ying, Kaipeng Zhang
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 |
|---|---|---|---|
| replace TSCM with spatial compression from Yume (ablation) | Yume1.5 · Yume-Bench | instruction following: 0.767 -> 0.836 | Improved |
| add TSCM instead of spatial compression (ablation) | Yume1.5 · Yume-Bench | motion smoothness: 0.973 -> 0.985 | No clear effect |
| add TSCM instead of spatial compression (ablation) | Yume1.5 · Yume-Bench | imaging quality: 0.733 -> 0.728 | No clear effect |
| add TSCM to bound history context (ablation) | Yume1.5 inference · autoregressive inference time vs video blocks | inference time: constant per-step time after 8 blocks; full context 12.41s at truncation | Improved |
| add Self-Forcing with TSCM training for few-step generation (ablation) | Yume1.5 · long-video T2V, 6th segment | aesthetic quality: 0.442 -> 0.523 | Improved |
| add Self-Forcing with TSCM training for few-step generation (ablation) | Yume1.5 · long-video T2V, 6th segment | imaging quality: 0.542 -> 0.601 | Improved |
| distill to 4 steps with Self-Forcing and TSCM, built on Wan2.2-5B | Yume1.5 · Yume-Bench | generation time (s): 107 -> 8 (Wan-2.2-5B vs Yume1.5) | Improved |
| switch from Yume to Yume1.5 | world generation · Yume-Bench | aesthetic quality: 0.518 -> 0.506 | Hurt |
| switch from Yume to Yume1.5 | world generation · Yume-Bench | imaging quality: 0.739 -> 0.728 | Hurt |
Later papers that compared with this method
2 later accepted papers report Yume 1.5 as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.
- Infinite-World: Scaling Interactive World Models to 1000-Frame Horizons via Pose-Free Hierarchical Memory (ICML 2026 · better on 3 of 4 shared results)
- Grounding World Simulation Models in a Real-World Metropolis (ECCV 2026 · 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
- SemanticDraw: Towards Real-Time Interactive Content Creation from Image Diffusion Models (CVPR 2025 · similarity 0.88)
- WorldMem: Long-term Consistent World Simulation with Memory (NeurIPS 2025 · similarity 0.86)
- Grid Diffusion Models for Text-to-Video Generation (CVPR 2024 · similarity 0.86)
- WorldWeaver: Generating Long-Horizon Video Worlds via Rich Perception (NeurIPS 2025 · similarity 0.86)
- Learning World Models for Interactive Video Generation (NeurIPS 2025 · similarity 0.86)
- GameGen-X: Interactive Open-world Game Video Generation (ICLR 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.
