MultihopSpatial: Multi-hop Compositional Spatial Reasoning Benchmark for Vision-Language Model
This addresses the problem of inadequate spatial reasoning benchmarks for Vision-Language Models, particularly for real-world deployment, though it is incremental as it builds on existing benchmark efforts.
The paper tackles the lack of benchmarks for multi-hop compositional spatial reasoning in Vision-Language Models by introducing MultihopSpatial, a benchmark with complex queries and a new metric (Acc@50IoU), and shows that current VLMs struggle with this challenge, with reinforcement learning post-training improving performance.
Spatial reasoning is foundational for Vision-Language Models (VLMs), particularly when deployed as Vision-Language-Action (VLA) agents in physical environments. However, existing benchmarks predominantly focus on elementary, single-hop relations, neglecting the multi-hop compositional reasoning and precise visual grounding essential for real-world scenarios. To address this, we introduce MultihopSpatial, offering three key contributions: (1) A comprehensive benchmark designed for multi-hop and compositional spatial reasoning, featuring 1- to 3-hop complex queries across diverse spatial perspectives. (2) Acc@50IoU, a complementary metric that simultaneously evaluates reasoning and visual grounding by requiring both answer selection and precise bounding box prediction - capabilities vital for robust VLA deployment. (3) MultihopSpatial-Train, a dedicated large-scale training corpus to foster spatial intelligence. Extensive evaluation of 37 state-of-the-art VLMs yields eight key insights, revealing that compositional spatial reasoning remains a formidable challenge. Finally, we demonstrate that reinforcement learning post-training on our corpus enhances both intrinsic VLM spatial reasoning and downstream embodied manipulation performance.