Vision-Language Memory for Spatial Reasoning
By: Zuntao Liu , Yi Du , Taimeng Fu and more
Potential Business Impact:
Robots understand 3D space better from videos.
Spatial reasoning is a critical capability for intelligent robots, yet current vision-language models (VLMs) still fall short of human-level performance in video-based spatial reasoning. This gap mainly stems from two challenges: a semantic-geometric misalignment that prevents consistent 3D understanding, and the absence of persistent memory to retain 3D representation and understanding over time. To address these limitations, we present VLM$^2$, a Vision-Language Model with persistent Memory for spatial reasoning with a view-consistent, 3D-aware representation purely from 2D video. Specifically, to enhance long-horizon reasoning, we incorporate a dual-memory module, consisting of a working memory that operates as a sliding window to focus on immediate context, and an episodic memory that consolidates and stores critical long-term information. This design enables efficient and long-horizon spatial reasoning with a fixed computational cost. Extensive experiments on multiple benchmarks show that VLM$^2$ achieves state-of-the-art performance among video-only models, significantly advancing the frontier of visual-spatial intelligence.
Similar Papers
Reasoning Path and Latent State Analysis for Multi-view Visual Spatial Reasoning: A Cognitive Science Perspective
Artificial Intelligence
Helps computers understand 3D spaces from different views.
G$^2$VLM: Geometry Grounded Vision Language Model with Unified 3D Reconstruction and Spatial Reasoning
CV and Pattern Recognition
Teaches computers to understand 3D space from pictures.
G$^2$VLM: Geometry Grounded Vision Language Model with Unified 3D Reconstruction and Spatial Reasoning
CV and Pattern Recognition
Helps computers understand 3D space from pictures.