发表机构
Purdue University; Rutgers University; University at Albany; Microsoft（普渡大学; 罗格斯大学; 奥尔巴尼大学; 微软）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
提出SPIRE框架，将页面级幻灯片个性化视为逆向规划问题，通过结构去噪和强化学习协作学习潜在设计意图，无需预设模板或用户详细指令。
AI 中文摘要
幻灯片设计需要个性化定制整体主题和页面布局。然而，当前基于AI智能体的方法难以处理细粒度的页面级设计。仅依赖预设模板或用户冗长指令，它们无法捕捉潜在的设计意图，导致页面级幻灯片个性化（PSP）问题未解决。为弥补这一差距，本文将PSP形式化为一个逆向规划问题。我们提出学习设计意图，而不假设使用任何特定执行工具（如PowerPoint、Beamer）的知识。然而，放弃对这些工具的控制使得问题难以端到端优化。为克服这一点，我们提出SPIRE，一个原则性框架来近似解决PSP。通过故意破坏干净幻灯片的视觉结构，SPIRE创建一个可验证的去噪任务，其中两个智能体通过强化学习（RL）学习协作改进可执行设计。我们证明了结构去噪是PSP的一致替代，并且多智能体公式严格降低了RL中的策略梯度方差。大量实验证明了SPIRE的优越性。
英文摘要
Slide design requires personalizing both deck themes and page layouts. Yet, current AI agent-based methods struggle with fine-grained, page-level design. Solely relying on prespecified templates or user verbose instructions, they fail to capture latent design intents, leaving Page-level Slide Personalization (PSP) unresolved. To close this gap, this work formulates PSP as an inverse planning problem. We propose to learn a design intent without assuming any knowledge of the specific executing tools (e.g., PowerPoint, Beamer) being used. However, relinquishing control over these tools makes the problem intractable to optimize end-to-end. To overcome this, we propose SPIRE, a principled framework to solve PSP approximately. By intentionally corrupting the visual structures of clean slides, SPIRE creates a verifiable task to denoise the corruption, whereby two agents learn to collaboratively refine executable designs via reinforcement learning (RL). We present a proof that structural denoising is a consistent surrogate for PSP, and that the multi-agent formulation strictly reduces policy gradient variance in RL. Extensive experiments demonstrate the superiority of SPIRE.
CommentsECCV 2026
