Relational Abstractions for Spatial Reasoning with Diffusion Models

2026-10-07

AI Summary

This study addresses the inherent limitations of diffusion models in satisfying structured spatial reasoning constraints and learning implicit logical rules. To this end, we propose a relation-knowledge-guided mechanism grounded in object-centric representations. Methodologically, our approach enhances the structured generation capabilities of diffusion models through unsupervised object discovery and relational abstraction, complemented by the construction of a large-scale spatial reasoning benchmark. Experimental results demonstrate that the proposed method significantly improves performance on complex reasoning tasks while achieving robust generalization to out-of-distribution scenarios.

Abstract

Diffusion models excel at image synthesis, but they remain limited in their ability to reliably satisfy structured spatial reasoning constraints. In conditional data distribution modeling tasks with implicit logical structure, such as puzzles defined by visible clues paired with consistent solutions, state-of-the-art generative models tend to approximate pixel-space distributions without learning the underlying logical rules required for inference. To address this limitation, we present a novel framework for spatial reasoning with diffusion models that leverages unsupervised object discovery and abstractions of object relations. We show that the relational knowledge derived from object-centric representations enriches diffusion models with structural primitives, allowing them to effectively guide the generative representation space during both training and inference, and enabling conditional image generation that satisfies reasoning constraints. Additionally, we introduce a large-scale generative spatial reasoning benchmark with four datasets inspired by human-solvable puzzles. Our results show that relational abstractions significantly improve reasoning capabilities of diffusion models on a variety of complex reasoning tasks, while enabling robust generalization in out-of-distribution settings.

Problem

• Diffusion Models
• Spatial Reasoning
• Relational Abstractions
• Conditional Image Generation
• Logical Constraints

Innovation

• Diffusion Models
• Spatial Reasoning
• Relational Abstractions
• Unsupervised Object Discovery
• Object-Centric Representations