Research map / Video / Video generation
Context Forcing: Consistent Autoregressive Video Generation with Long Context
Shuo Chen, Cong Wei, Sun Sun, Tiancheng SHEN, Ping Nie, Kai Zou, Ge Zhang, Ming-Hsuan Yang, Wenhu Chen
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 contextual distribution matching distillation (ablation) | Context Forcing · VBench 60s video | total score: 80.36 -> 82.45 | Improved |
| add contextual distribution matching distillation (ablation) | Context Forcing · VBench 60s video | semantic score: 72.70 -> 76.10 | Improved |
| add contextual distribution matching distillation (ablation) | Context Forcing · VBench 60s video | dynamic degree: 48.12 -> 58.26 | Improved |
| add bounded positional encoding (ablation) | Context Forcing · VBench 60s video | total score: 73.52 -> 82.45 | Improved |
| add bounded positional encoding (ablation) | Context Forcing · VBench 60s video | subject consistency: 79.24 -> 94.88 | Improved |
| add bounded positional encoding (ablation) | Context Forcing · VBench 60s video | background consistency: 84.68 -> 95.34 | Improved |
| use similarity-based slow memory selection instead of uniform interval 1 (ablation) | Context Forcing · VBench 60s video | total score: 80.82 -> 82.45 | Improved |
| use similarity-based slow memory selection instead of uniform interval 2 (ablation) | Context Forcing · VBench 60s video | total score: 81.11 -> 82.45 | Improved |
| use similarity-based slow memory selection instead of uniform interval 1 (ablation) | Context Forcing · VBench 60s video | semantic score: 75.32 -> 76.10 | No clear effect |
| Context Forcing student versus LongLive on 60s DINO | Context Forcing · Single-prompt 60s consistency | DINO score at 60s: 86.26 -> 87.89 | Improved |
| Context Forcing student versus LongLive on 60s VBench total | Context Forcing · VBench 60s video | total score: 83.64 -> 82.45 | Hurt |
| Context Forcing student versus LongLive on 5s VBench total | Context Forcing · VBench 5s video | total score: 84.87 -> 83.44 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.
- LongLive — LongLive: Real-time Interactive Long Video Generation (ICLR 2026) · better on 4 of 6 shared results
- Infinity-RoPE — Infinity-RoPE: Action-Controllable Infinite Video Generation Emerges From Autoregressive Self-Rollout (CVPR 2026) · better on 5 of 5 shared results
- Rolling Forcing — Rolling Forcing: Autoregressive Long Video Diffusion in Real Time (ICLR 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
- Self-Forcing++: Towards Minute-Scale High-Quality Video Generation (ICLR 2026 · similarity 0.92)
- Learning-to-Memorize: Dynamic Context Management for Long-Horizon Autoregressive Video Generation (NeurIPS 2026 · similarity 0.90)
- Taming Teacher Forcing for Masked Autoregressive Video Generation (CVPR 2025 · similarity 0.90)
- Deep Forcing: Training-Free Long Video Generation with Deep Sink and Participative Compression (ICML 2026 · similarity 0.90)
- Ring Forcing: Towards Precise Long-Term Memory for Autoregressive Video Diffusion (ECCV 2026 · similarity 0.89)
- Self Forcing: Bridging the Train-Test Gap in Autoregressive Video Diffusion (NeurIPS 2025 · similarity 0.87)
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.
