{"ID":5551927,"CreatedAt":"2026-07-02T01:54:51.863792489Z","UpdatedAt":"2026-07-04T03:57:10.279025113Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2607.00407","arxiv_id":"2607.00407","title":"Personalization as Inverse Planning: Learning Latent Design Intents for Agentic Slide Generation via Structural Denoising","abstract":"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.","short_abstract":"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) unre...","url_abs":"https://arxiv.org/abs/2607.00407","url_pdf":"https://arxiv.org/pdf/2607.00407v1","authors":"[\"Tianci Liu\",\"Zihan Dong\",\"Linjun Zhang\",\"Haoyu Wang\",\"jing Gao\",\"Emre Kiciman\",\"Ranveer Chandra\",\"Wei-Ting Chen\"]","published":"2026-07-01T04:05:47Z","proceeding":"cs.AI","tasks":"[\"cs.AI\"]","methods":"[\"Reinforcement Learning\"]","has_code":false}