CorrFlow paper adds a lightweight refiner that self-corrects Flow Matching, boosting image and video generation
udmrzn · x · 2026-10-20
- New arXiv paper CorrFlow introduces self-correcting flow matching: a lightweight refiner reuses intermediate features and predicted clean samples to correct the base model's velocity predictions each step.
- Theoretically motivated: under rectified-flow interpolation, one-step clean-prediction error equals the realized velocity residual scaled by noise level; the paper proves conditions where feature reuse beats affine predictors and refiner training yields greater residual-risk reduction than continued base-model training at matched compute.
- Adds no extra base-model evaluations—just one lightweight refiner pass per step.
- Consistent gains in visual quality and physical consistency on class-conditional image generation, text-to-video, and image-to-video.
More from Research
- Laude Institute launches FrontierCS 2: agents must beat deleted paper contributions blind — lateinteraction · 2026-10-23
- RFD4-Proteina's 'underrated' docs detail partial diffusion for protein redesign — AllThingsApx · 2026-10-23
- MPI-Sec PhD Thesis on Vulnerability Detection Benchmarks Includes Pre-LLM Classic "Top Score on the Wrong Exam" — mboehme_ · 2026-10-23
- LiveNerf's 29-day data finds no nerfing of Claude Opus 5.5 — TheOnlyVibemaster · 2026-10-23
- New paper makes 9 falsifiable predictions toward a major AI-enabled terror attack by 2027-2028 — StephenLCasper · 2026-10-23
- HumanoidPose recovers humanoid robots' full-body joint states from RGB video alone — cholz · 2026-10-23