arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2511.18352cs.CV
收藏

MagicWand: 一个通用代理用于生成与评估,与用户偏好对齐

MagicWand: A Universal Agent for Generation and Evaluation Aligned with User Preference

Shanghai Jiao Tong University(上海交通大学);De Artificial Intelligence Lab(德人工智能实验室);The Chinese University of Hong Kong, Shenzhen China(香港中文大学(深圳))

Zitong Xu, Dake Shen, Yaosong Du, Kexiang Hao, Jinghan Huang, Xiande Huang

AI总结:

MagicWand通过结合用户偏好数据集和先进生成模型,实现高质量内容生成与偏好对齐评估。

摘要:

近年来,AIGC(人工智能生成内容)模型的进步在图像和视频生成方面取得了显著进展。然而,用户仍然难以获得符合其偏好的内容,因为构建详细提示困难且缺乏保留其偏好的机制。为了解决这些挑战,我们构建了UniPrefer-100K,一个包含图像、视频和关联文本的大规模数据集,这些文本描述了用户倾向于偏好的风格。基于UniPrefer-100K,我们提出了MagicWand,一个通用的生成与评估代理,它根据用户偏好增强提示,利用先进的生成模型生成高质量内容,并应用与偏好对齐的评估和优化。此外,我们引入了UniPreferBench,第一个大规模基准,拥有超过120,000个注释,用于评估用户偏好对齐在多样化AIGC任务中的表现。在UniPreferBench上的实验表明,MagicWand在各种场景中都能一致地生成和评估与用户偏好高度一致的内容。

英文摘要:

Recent advances in AIGC (Artificial Intelligence Generated Content) models have enabled significant progress in image and video generation. However, users still struggle to obtain content that aligns with their preferences due to the difficulty of crafting detailed prompts and the lack of mechanisms to retain their preferences. To address these challenges, we construct \textbf{UniPrefer-100K}, a large-scale dataset comprising images, videos, and associated text that describes the styles users tend to prefer. Based on UniPrefer-100K, we propose \textbf{MagicWand}, a universal generation and evaluation agent that enhances prompts based on user preferences, leverages advanced generation models for high-quality content, and applies preference-aligned evaluation and refinement. In addition, we introduce \textbf{UniPreferBench}, the first large-scale benchmark with over 120K annotations for assessing user preference alignment across diverse AIGC tasks. Experiments on UniPreferBench demonstrate that MagicWand consistently generates content and evaluations that are well aligned with user preferences across a wide range of scenarios.

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。

↑