AutoLayout: Closed-Loop Layout Synthesis via Slow-Fast Collaborative Reasoning

2025-07-06

AI Summary

Automated layout generation suffers from severe spatial hallucination and struggles to simultaneously ensure semantic fidelity and physical plausibility, often resulting in floating, overlapping, or topologically inconsistent objects. To address this, we propose a slow-fast collaborative reasoning framework: the slow system performs fine-grained spatial and semantic reasoning, while the fast system efficiently generates discrete coordinates and topological relations. We introduce an LLM-based adaptive relation repository for rule-free semantic modeling, and design a Reasoning-Reflection-Generation (RRG) closed-loop self-validation mechanism enabling iterative error correction. Evaluated across eight real-world scenarios, our method achieves a 10.1% improvement in the composite metric—encompassing physical plausibility, semantic consistency, and functional completeness—over state-of-the-art approaches, significantly mitigating spatial hallucination.

Technology Category

• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning
• Planning, Routing, and Scheduling: Model-Based Reasoning
• Computer Vision: Visual Reasoning & Symbolic Representations

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

The automated generation of layouts is vital for embodied intelligence and autonomous systems, supporting applications from virtual environment construction to home robot deployment. Current approaches, however, suffer from spatial hallucination and struggle with balancing semantic fidelity and physical plausibility, often producing layouts with deficits such as floating or overlapping objects and misaligned stacking relation. In this paper, we propose AutoLayout, a fully automated method that integrates a closed-loop self-validation process within a dual-system framework. Specifically, a slow system harnesses detailed reasoning with a Reasoning-Reflection-Generation (RRG) pipeline to extract object attributes and spatial constraints. Then, a fast system generates discrete coordinate sets and a topological relation set that are jointly validated. To mitigate the limitations of handcrafted rules, we further introduce an LLM-based Adaptive Relation Library (ARL) for generating and evaluating layouts. Through the implementation of Slow-Fast Collaborative Reasoning, the AutoLayout efficiently generates layouts after thorough deliberation, effectively mitigating spatial hallucination. Its self-validation mechanism establishes a closed-loop process that iteratively corrects potential errors, achieving a balance between physical stability and semantic consistency. The effectiveness of AutoLayout was validated across 8 distinct scenarios, where it demonstrated a significant 10.1% improvement over SOTA methods in terms of physical plausibility, semantic consistency, and functional completeness.

Problem

• Automated layout generation with spatial hallucination issues
• Balancing semantic fidelity and physical plausibility in layouts
• Correcting floating, overlapping objects and misaligned stacking relations

Innovation

• Closed-loop self-validation process integration
• Slow-fast collaborative reasoning framework
• LLM-based Adaptive Relation Library