TailSteer: Trajectory-Tail Steering for Drift-Resistant Multi-turn Image Editing
[Widzenie komputerowe] Autorzy: Ruikang Li, Zixuan Chen, Jiarui Wu i 6 więcej
Strong single-turn editing performance does not guarantee reliable multi-turn interaction: as edited outputs are repeatedly reused, small unintended changes can accumulate into severe visual drift. This long-term drift is often mitigated through training on multi-turn data or cross-turn rollouts, potentially altering pretrained priors, or through specialized inference-time mechanisms that are difficult to scale to foundation models. We introduce TailSteer, a final-stage training framework that improves multi-turn robustness while retaining the generative and editing capabilities acquired through earlier training. TailSteer follows a simple principle: keep early denoising dynamics close to those of the base model and concentrate preservation corrections near the trajectory tail, where fine-grained deviations can be suppressed. These corrections are learned from the editor's own on-policy…
- Kategoria arxiv
- cs.CV
- Kategoria
- Widzenie komputerowe