yyuncong/MindJourney-World-Model

<br/> <h1 align="center" style="font-size: 1.7rem">MindJourney: Test-Time Scaling with World Models for Spatial Reasoning</h1> <p align="center"> NeurIPS 2025 </p> <p align="center"> <a href="https://yyuncong.github.io/">Yuncong Yang</a>, <a href="https://jiagengliu02.github.io/">Jiageng Liu</a>, <a href="https://cozheyuanzhangde.github.io/">Zheyuan Zhang</a>, <a href="https://rainbow979.github.io/">Siyuan Zhou</a>, <a href="https://cs-people.bu.edu/rxtan/">Reuben Tan</a>, <a href="https://jwyang.github.io/">Jianwei Yang</a>, <a href="https://yilundu.github.io/">Yilun Du</a>, <a href="https://people.csail.mit.edu/ganchuang">Chuang Gan</a> </p> <p align="center"> <a href="https://arxiv.org/abs/2507.12508"> <img src='https://img.shields.io/badge/Paper-PDF-red?style=flat&logo=arXiv&logoColor=red' alt='Paper PDF'> </a> <a href='https://umass-embodied-agi.github.io/MindJourney/' style='padding-left: 0.5rem;'> <img src='https://img.shields.io/badge/Project-Page-blue?style=flat&logo=Google%20chrome&logoColor=blue' alt='Project Page'> </a> </p>

</p>

MindJourney is a test-time scaling framework that leverages the 3D imagination capability of World Models to strengthen spatial reasoning in Vision-Language Models (VLMs). We evaluate on the SAT dataset and provide a baseline pipeline, a Stable Virtual Camera (SVC) based spatial beam search pipeline, and a Search World Model (SWM) based spatial beam search pipeline.