可解释性潜在奖励的自校正图像生成
Self-Corrected Image Generation with Explainable Latent Rewards
AI总结：
本文提出xLARD框架，利用多模态大语言模型通过可解释性潜在奖励指导生成，改进语义对齐和视觉保真度，同时保持生成先验。
摘要：
尽管文本到图像生成取得显著进展，但将输出与复杂提示对齐仍具挑战，尤其在细粒度语义和空间关系方面。这种困难源于生成的前馈性质，要求在完全理解输出前预测对齐。相反，评估生成图像更具可操作性。受此不对称性启发，我们提出xLARD，一种自校正框架，利用多模态大语言模型通过可解释性潜在奖励指导生成。xLARD引入轻量级校正器，基于模型生成的参考的结构化反馈细化潜在表示。关键组件是可微映射从潜在编辑到可解释奖励信号，使模型能从非可微图像级评估中持续获得潜在层面指导。此机制使模型在生成过程中能理解、评估和纠正自身。在多样化的生成和编辑任务中实验表明，xLARD在提高语义对齐和视觉保真度的同时保持生成先验。代码可在https://yinyiluo.github.io/xLARD/获取。
英文摘要：
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. Code is available at https://yinyiluo.github.io/xLARD/.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
