基于可解释潜变量奖励的自纠正图像生成
Self-Corrected Image Generation with Explainable Latent Rewards
把不可微的图像级评价（MLLM 打分）通过可微映射转成潜变量空间的连续奖励梯度，再用轻量 corrector 在采样过程中修正潜变量——这套「潜变量奖励 + 自纠正」机制可直接迁移到任何 diffusion/flow 采样器的对齐微调或测试时优化中。
原文依据 a differentiable mapping from latent edits to interpretable reward signals, enabling continuous latent-level guidance from non-differentiable image-level evaluations
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论尽管文本到图像生成已取得显著进展，但使输出与复杂提示对齐仍然困难，尤其是细粒度语义和空间关系方面。这一困难源于生成的前馈特性：模型必须在尚未充分理解输出的情况下预先设想对齐。相比之下，评估已生成的图像则更为可行。受这种不对称性的启发，我们提出 xLARD，一个利用多模态大语言模型、通过可解释潜变量奖励来引导生成的自纠正框架。xLARD 引入一个轻量级纠正器，依据模型生成参考所提供的结构化反馈来精炼潜变量表示。其关键组件是一个从潜变量编辑到可解释奖励信号的可微映射，使得来自不可微图像级评估的连续潜变量级引导成为可能。该机制让模型在生成过程中理解、评估并纠正自身。在多种生成与编辑任务上的实验表明，xLARD 在保持生成先验的同时提升了语义对齐与视觉保真度，为弥合理解与合成之间的差距提供了一种数据高效且可泛化的方案。
英文原文摘要 1242 字符
Despite significant progress in text-to-image generation, aligning outputs with complex prompts remains challenging, particularly for fine-grained semantics and spatial relations. This difficulty stems from the feed-forward nature of generation, which requires anticipating alignment without fully understanding the output. In contrast, evaluating generated images is more tractable. Motivated by this asymmetry, we propose xLARD, a self-correcting framework that uses multimodal large language models to guide generation through Explainable Latent Rewards. xLARD introduces a lightweight corrector that refines latent representations based on structured feedback from model-generated references. A key component is a differentiable mapping from latent edits to interpretable reward signals, enabling continuous latent-level guidance from non-differentiable image-level evaluations. This mechanism allows the model to understand, assess, and correct itself during generation. Experiments across diverse generation and editing tasks show that xLARD improves semantic alignment and visual fidelity while maintaining generative priors, offering a data-efficient and generalizable solution to bridging the gap between comprehension and synthesis.