MindJourney: Test-Time Scaling with World Models for Spatial Reasoning
This addresses a critical limitation in VLMs for embodied AI tasks like navigation and manipulation, offering a plug-and-play solution for robust 3D reasoning.
The paper tackles the problem of spatial reasoning in 3D space for vision-language models (VLMs), which struggle with tasks like anticipating scene changes after motion, by proposing MindJourney, a test-time scaling framework that couples VLMs with a controllable world model based on video diffusion, resulting in a 7.7% performance boost on the SAT benchmark without fine-tuning.
Spatial reasoning in 3D space is central to human cognition and indispensable for embodied tasks such as navigation and manipulation. However, state-of-the-art vision-language models (VLMs) struggle frequently with tasks as simple as anticipating how a scene will look after an egocentric motion: they perceive 2D images but lack an internal model of 3D dynamics. We therefore propose MindJourney, a test-time scaling framework that grants a VLM with this missing capability by coupling it to a controllable world model based on video diffusion. The VLM iteratively sketches a concise camera trajectory, while the world model synthesizes the corresponding view at each step. The VLM then reasons over this multi-view evidence gathered during the interactive exploration. Without any fine-tuning, our MindJourney achieves over an average 7.7% performance boost on the representative spatial reasoning benchmark SAT, showing that pairing VLMs with world models for test-time scaling offers a simple, plug-and-play route to robust 3D reasoning. Meanwhile, our method also improves upon the test-time inference VLMs trained through reinforcement learning, which demonstrates the potential of our method that utilizes world models for test-time scaling.