基础编码器即可实现偏好感知的个性化
Foundation Encoders Are All You Need for Preference-Aware Personalization
把个性化偏好注入完全放在基础编码器（文本编码器/CLIP 编码器）的表征空间里完成，不改扩散主干、不做微调，为个性化生成提供了一条可跨多个 T2I 底座复用的轻量路线；做 personalization 或 adapter 设计时可直接借鉴其「重构 transformer 编码器以融合偏好」的做法。
原文依据 reconstructs transformer-based encoders to integrate them while preserving target fidelity
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论基于用户行为的个性化图像生成能够在极少用户干预下反映个体偏好。然而，现有研究往往依赖不准确的用户画像、高昂的资源成本以及针对特定模型的设计，这些因素共同限制了创造性、多样性与通用性。为解决上述局限，我们提出 FAN，一种仅使用基础编码器、无需额外结构即可实现偏好感知个性化的新方法。FAN 通过定制化画像捕捉用户偏好，并重构基于 Transformer 的编码器以将其融入，同时保持目标保真度。实验表明，FAN 在多种基础文本到图像模型上均能实现稳健的个性化。它还可扩展至 CLIP 检索、unCLIP 以及视觉语言模型等应用，无需微调即可无缝集成到多种编码器中。
英文原文摘要 881 字符
Personalized image generation based on user behavior reflects individual preferences with minimal user intervention. However, existing studies often rely on inaccurate profiling, high resource costs, and model-specific designs, which jointly restrict creativity, diversity, and generality. To address these limitations, we propose FAN, a novel approach that enables preference-aware personalization using only foundation encoders, without additional structures. FAN performs tailored profiling to capture user preferences and reconstructs transformer-based encoders to integrate them while preserving target fidelity. Experiments show that FAN achieves robust personalization across various foundation text-to-image models. It also extends to applications such as CLIP retrieval, unCLIP, and vision-language models, seamlessly integrating into diverse encoders without fine-tuning.