Flow-of-Thought: A Framework for Visual Reasoning

2026-10-07

AI Summary

This study addresses the limited spatial understanding capabilities of large language models by proposing an interpretable visual reasoning framework. The method introduces visual sketches as intermediate reasoning steps, simulating human mental imagery to enhance spatial cognition. It pioneers the integration of coordinate-aware trajectory flow fields with generative hypothesis comparison, employing flow field training on SO(2) group orbits alongside foreground-weighted reconstruction energy contrast techniques. Experimental results demonstrate that the proposed framework achieves 100% and 99% accuracy on Tetris and colored-shape benchmarks, respectively, significantly improving multi-view reasoning performance.

Abstract

Mental imagery, ``seeing with the mind's eye'' is an essential aspect of human cognition. Despite rapid progress Large Language Models (LLMs) and Vision Transformers (ViTs) still underperform on tasks requiring spatial understanding. To address this, we introduce Flow-of-Thought (FoT), a framework that integrates the generation of visual sketches as intermediate reasoning steps, mimicking mental imagery in humans. We train coordinate-aware trajectory flow fields on $SO(2)$ group orbits and cumulative shortest paths, then freeze the learned dynamics; same vs. different decisions compare competing generative hypotheses using foreground-weighted reconstruction energy. On locked tests FoT reaches 100.0% accuracy on Tetris and 99.0% on colored shapes. Under frozen transfer, the orbit-trained 2D flow improves over its endpoint-only control on BLINK Multi-view (72.2% vs. 63.9% on 133 public validation pairs), supporting continuous visual traces as an effective and interpretable representation for spatial reasoning in some out-of-distribution settings.

Problem

• visual reasoning
• spatial understanding
• mental imagery
• Large Language Models
• Vision Transformers

Innovation

• Flow-of-Thought
• Visual Reasoning
• Mental Imagery
• Trajectory Flow Fields
• Spatial Understanding