SVQA-R1: Reinforcing Spatial Reasoning in MLLMs via View-Consistent Reward Optimization

TL;DR

This paper introduces SVQA-R1, a novel reinforcement learning framework that enhances spatial reasoning in vision-language models for VQA tasks by using view-consistent rewards, leading to improved accuracy and interpretability.

Contribution

We extend the R1 paradigm to spatial VQA with a new group-wise RL strategy called Spatial-GRPO, promoting grounded spatial understanding without supervised fine-tuning.

Findings

Significant accuracy improvements on spatial VQA benchmarks.

Model exhibits interpretable reasoning paths.

Effective across multiple spatial reasoning tasks.

Abstract

Spatial reasoning remains a critical yet underdeveloped capability in existing vision-language models (VLMs), especially for Spatial Visual Question Answering (Spatial VQA) tasks that require understanding relative positions, distances, and object configurations. Inspired by the R1 paradigm introduced in DeepSeek-R1, which enhances reasoning in language models through rule-based reinforcement learning (RL), we propose SVQA-R1, the first framework to extend R1-style training to spatial VQA. In particular, we introduce Spatial-GRPO, a novel group-wise RL strategy that constructs view-consistent rewards by perturbing spatial relations between objects, e.g., mirror flipping, thereby encouraging the model to develop a consistent and grounded understanding of space. Our model, SVQA-R1, not only achieves dramatically improved accuracy on spatial VQA benchmarks but also exhibits interpretable reasoning paths even without using supervised fine-tuning (SFT) data. Extensive experiments and visualization demonstrate the effectiveness of SVQA-R1 across multiple spatial reasoning benchmarks.