AI修图连改20轮也不跑偏
多轮修图真正的麻烦不是单次改得丑，而是改到第十几轮时画面已经悄悄面目全非——每次微小的意外改动会累积成严重漂移。这篇的做法很克制：不动模型早段去噪的走向(保住原有能力)，只在生成轨迹的尾段集中做「保真修正」，而且这些修正不用人工标注的成对数据，直接拿「没被改动的原内容」当参照，从模型自己的去噪过程里学。在 FLUX、SenseNova、Qwen 三个系列上，连改 15-20 轮累积漂移明显变小，同时文生图和单图编辑的原有水平基本没掉。它不是你明天能直接用的产品，但意味着「让 AI 陪你反复改一张图而不崩」这件事，开始能在基础模型上靠最后一程训练解决，而不是逼模型重学一遍。
📄 原文摘要(英文)
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 denoising rollouts, using unchanged source content as a preservation reference without requiring curated ground-truth editing pairs. Experiments across the FLUX, SenseNova, and Qwen families show that TailSteer substantially reduces cumulative drift over 15-20 editing turns while largely preserving base-model performance on extensive text-to-image and single-image editing benchmarks. These results demonstrate that long-horizon editing robustness can be unlocked in foundation models after TailSteer, rather than by relearning their existing capabilities.