Research map / Video / Video generation
Generative Video Bi-flow
Chen Liu, Tobias Ritschel
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 bilinear extension (joint time and denoising flow field) (ablation) | video streaming flow · SKY | FVD: 2144 -> 526 | Improved |
| add bilinear extension (joint time and denoising flow field) (ablation) | video streaming flow · BIKING | FVD: 14010 -> 3753 | Improved |
| add bilinear extension (ablation) | video streaming flow · SKY | ODE steps per frame: 7.84 -> 14.57 | Hurt |
| replace conditional diffusion with Bi-flow | ConDiff · SKY | FVD: 1271 -> 526 | Improved |
| replace conditional diffusion with Bi-flow | ConDiff · CARLA | ODE steps per frame: 53.77 -> 23.51 | Improved |
| replace conditional diffusion with Bi-flow | ConDiff · MAZES | FVD: 424 -> 605 | Hurt |
| replace conditional diffusion with Bi-flow under 20 equal steps | ConDiff · SKY | FVD: 994 -> 491 | Improved |
| replace conditional diffusion with Bi-flow under 20 equal steps | ConDiff · MAZES | FVD: 655 -> 620 | Improved |
| increase sampling noise level from 0.0 to 0.3 (ablation) | Bi-flow · long-video generation | FVD slope over time | Improved |
| increase sampling noise level (ablation) | Bi-flow · average FVD | FVD | Mixed |
Closest papers by meaning
- From Slow Bidirectional to Fast Autoregressive Video Diffusion Models (CVPR 2025 · similarity 0.88)
- Generative Pre-trained Autoregressive Diffusion Transformer (NeurIPS 2025 · similarity 0.88)
- BIVDiff: A Training-Free Framework for General-Purpose Video Synthesis via Bridging Image and Video Diffusion Models (CVPR 2024 · similarity 0.88)
- Pyramidal Flow Matching for Efficient Video Generative Modeling (ICLR 2025 · similarity 0.88)
- Equilibrium Forcing: Adaptive Video Generation Without Noise Conditioning (NeurIPS 2026 · similarity 0.87)
- FreeNoise: Tuning-Free Longer Video Diffusion via Noise Rescheduling (ICLR 2024 · 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.
