Research map / Video / Video generation
VOGUE: Unified Understanding, Generation, and Editing for Videos
Cong Wei, Quande Liu, Zixuan Ye, Qiulin Wang, Xintao Wang, Pengfei Wan, Kun Gai, 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 joint multi-task training over single-task models (ablation) | UniVideo · in-context tasks average | prompt following: 0.64 -> 0.80 | Improved |
| add joint multi-task training over single-task models (ablation) | UniVideo · in-context tasks average | subject consistency: 0.67 -> 0.78 | Improved |
| add joint multi-task training over single-task models (ablation) | UniVideo · in-context tasks average | video quality: 0.79 -> 0.85 | Improved |
| add joint multi-task training over single-task models (ablation) | UniVideo · in-context swap | prompt following: 0.53 -> 0.91 | Improved |
| add joint multi-task training over single-task models (ablation) | UniVideo · in-context multi-id generation | prompt following: 0.75 -> 0.75 | No clear effect |
| feed visual inputs to MMDiT as well as MLLM (ablation) | UniVideo · in-context tasks average | subject consistency: 0.18 -> 0.78 | Improved |
| feed visual inputs to MMDiT as well as MLLM (ablation) | UniVideo · in-context tasks average | prompt following: 0.66 -> 0.80 | Improved |
| feed visual inputs to MMDiT as well as MLLM (ablation) | UniVideo · in-context multi-id generation | prompt following: 0.81 -> 0.75 | Hurt |
| use MMDiT self-attention conditioning over cross-attention DiT (ablation) | UniVideo · video generation (15K steps) | prompt alignment | Improved |
| freeze the MLLM for understanding | UniVideo · MMBench | accuracy: 83.5 | No clear effect |
Closest papers by meaning
- MME-Unify: A Comprehensive Benchmark for Unified Multimodal Understanding and Generation Models (ICLR 2026 · similarity 0.89)
- DuoGen: Towards Autonomous Interleaved Multimodal Generation (CVPR 2026 · similarity 0.88)
- Guiding Instruction-based Image Editing via Multimodal Large Language Models (ICLR 2024 · similarity 0.88)
- OmniGen2: Towards Instruction-Aligned Multimodal Generation (CVPR 2026 · similarity 0.88)
- VicEdit: Learning to Edit Videos from Visual In-Context Examples (NeurIPS 2026 · similarity 0.88)
- Text-Vision Co-Instructed Image Editing (NeurIPS 2026 · 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.
