{"ID":2866256,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2509.21662","arxiv_id":"2509.21662","title":"MMPlanner: Zero-Shot Multimodal Procedural Planning with Chain-of-Thought Object State Reasoning","abstract":"Multimodal Procedural Planning (MPP) aims to generate step-by-step instructions that combine text and images, with the central challenge of preserving object-state consistency across modalities while producing informative plans. Existing approaches often leverage large language models (LLMs) to refine textual steps; however, visual object-state alignment and systematic evaluation are largely underexplored. We present MMPlanner, a zero-shot MPP framework that introduces Object State Reasoning Chain-of-Thought (OSR-CoT) prompting to explicitly model object-state transitions and generate accurate multimodal plans. To assess plan quality, we design LLM-as-a-judge protocols for planning accuracy and cross-modal alignment, and further propose a visual step-reordering task to measure temporal coherence. Experiments on RECIPEPLAN and WIKIPLAN show that MMPlanner achieves state-of-the-art performance, improving textual planning by +6.8%, cross-modal alignment by +11.9%, and visual step ordering by +26.7%","short_abstract":"Multimodal Procedural Planning (MPP) aims to generate step-by-step instructions that combine text and images, with the central challenge of preserving object-state consistency across modalities while producing informative plans. Existing approaches often leverage large language models (LLMs) to refine textual steps; ho...","url_abs":"https://arxiv.org/abs/2509.21662","url_pdf":"https://arxiv.org/pdf/2509.21662v1","authors":"[\"Afrina Tabassum\",\"Bin Guo\",\"Xiyao Ma\",\"Hoda Eldardiry\",\"Ismini Lourentzou\"]","published":"2025-09-25T22:31:19Z","proceeding":"cs.LG","tasks":"[\"cs.LG\"]","methods":"[\"Large Language Model\",\"Language Model\"]","has_code":false}