From Diagnosis to Improvement: Probing Spatio-Physical Reasoning in Vision Language Models
By: Tiancheng Han , Yunfei Gao , Yong Li and more
Potential Business Impact:
Teaches computers to understand how things move.
Spatio-physical reasoning, a foundation capability for understanding the real physics world, is a critical step towards building robust world models. While recent vision language models (VLMs) have shown remarkable progress in specialized domains like multimodal mathematics and pure spatial understanding, their capability for spatio-physical reasoning remains largely unexplored. This paper provides a comprehensive diagnostic analysis of mainstream VLMs, revealing that current models perform inadequately on this crucial task. Further detailed analysis shows that this underperformance is largely attributable to biases caused by human-like prior and a lack of deep reasoning. To address these challenges, we apply supervised fine-tuning followed by rule-based reinforcement learning to Qwen2.5-VL-7B, resulting in significant improvements in spatio-physical reasoning capabilities and surpassing leading proprietary models. Nevertheless, despite this success, the model's generalization to new physics scenarios remains limited -- underscoring the pressing need for new approaches in spatio-physical reasoning.
Similar Papers
Vision-Language Memory for Spatial Reasoning
CV and Pattern Recognition
Robots understand 3D space better from videos.
Interpretable Physics Reasoning and Performance Taxonomy in Vision-Language Models
Machine Learning (CS)
Tests if computers understand how things move.
SpaRE: Enhancing Spatial Reasoning in Vision-Language Models with Synthetic Data
CV and Pattern Recognition
Teaches computers to understand where things are.