From Pixels to Policies: Reinforcing Spatial Reasoning in Language Models for Content-Aware Layout Design

• URL: http://arxiv.org/abs/2602.13912v2
• Date: Tue, 17 Feb 2026 22:24:20 GMT
• Title: From Pixels to Policies: Reinforcing Spatial Reasoning in Language Models for Content-Aware Layout Design
• Abstract summary: LaySPA is a reinforcement learning framework for content-aware graphic layout design.<n>Instead of operating at the pixel level, we reformulate layout design as a policy learning problem over a structured textual spatial environment.<n>LaySPA produces dual-level outputs comprising interpretable reasoning traces and structured layout specifications.
• Score: 21.929198075025187
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: We introduce LaySPA, a reinforcement learning framework that equips large language models (LLMs) with explicit and interpretable spatial reasoning for content-aware graphic layout design. LaySPA addresses two key challenges: LLMs' limited spatial reasoning and the lack of opacity in design decision making. Instead of operating at the pixel level, we reformulate layout design as a policy learning problem over a structured textual spatial environment that explicitly encodes canvas geometry, element attributes, and inter-element relationships. LaySPA produces dual-level outputs comprising interpretable reasoning traces and structured layout specifications, enabling transparent and controllable design decision making. Layout design policy is optimized via a multi-objective spatial critique that decomposes layout quality into geometric validity, relational coherence, and aesthetic consistency, and is trained using relative group optimization to stabilize learning in open-ended design spaces. Experiments demonstrate that LaySPA improves structural validity and visual quality, outperforming larger proprietary LLMs and achieving performance comparable to specialized SOTA layout generators while requiring fewer annotated samples and reduced latency.

Related papers

• Learning an Interior Layout Policy in a Domain Specific Language Action Space [4.7288373059475814]
  Existing methods often oversimplify the task by reducing room conditions to coarse 3D bounding boxes and neglecting structural elements such as doors and windows.<n>We propose a novel framework for learning an interior layout policy in a domain-specific language () action space.
  arXiv  Detail & Related papers  (2026-07-31T09:28:19Z)
• SpaceEra++: A Unified Framework Towards 3D Spatial Reasoning in Video [88.79625979053873]
  We propose a novel framework, SpaceEra, for pre-trained vision-language models (VLMs)<n>We extend the original framework into a comprehensive system, termed SpaceEra++, which spans data construction, model design, training optimization, and prompting inference.<n>In addition, to enhance spatial reasoning, we develop SpaceAlign, which enforces pairwise object constraints by jointly exploiting absolute coordinates and relative spatial relations.
  arXiv  Detail & Related papers  (2026-07-02T06:56:29Z)
• Design-MLLM: A Reinforcement Alignment Framework for Verifiable and Aesthetic Interior Design [23.825603561362627]
  Interior design is a requirements-to-visual-plan generation process.<n> MLLMs produce layouts that are unbuildable and aesthetically inconsistent.<n>We propose Design-MLLM, a reinforcement alignment framework.
  arXiv  Detail & Related papers  (2026-03-04T06:51:28Z)
• Co-Layout: LLM-driven Co-optimization for Interior Layout [8.182031753612875]
  We present a framework for automated interior design that combines large language models (LLMs) with grid-based integer programming to jointly optimize room layout and furniture placement.<n>Our formulation accounts for key design requirements, including corridor connectivity, room accessibility, spatial exclusivity, and user-specified preferences.
  arXiv  Detail & Related papers  (2025-11-16T06:20:55Z)
• DisCo-Layout: Disentangling and Coordinating Semantic and Physical Refinement in a Multi-Agent Framework for 3D Indoor Layout Synthesis [76.7196710324494]
  3D indoor layout synthesis is crucial for creating virtual environments.<n>DisCo is a novel framework that disentangles and coordinates physical and semantic refinement.
  arXiv  Detail & Related papers  (2025-10-02T16:30:37Z)
• Light-SQ: Structure-aware Shape Abstraction with Superquadrics for Generated Meshes [60.92139345612904]
  We present Light-SQ, a novel superquadric-based optimization framework.<n>We propose a block-regrow-fill strategy guided by structure-aware volumetric decomposition.<n>Experiments demonstrate that Light-SQ enables efficient, high-fidelity, and editable shape abstraction with superquadrics.
  arXiv  Detail & Related papers  (2025-09-29T16:18:32Z)
• LLMs as Layout Designers: Enhanced Spatial Reasoning for Content-Aware Layout Generation [27.377286708795506]
  LaySPA is a reinforcement learning-based framework that augments Large Language Models with explicit spatial reasoning capabilities for layout design.<n>We show that LaySPA substantially improves the generation of structurally valid and visually appealing layouts.
  arXiv  Detail & Related papers  (2025-09-21T03:02:59Z)
• ReLayout: Integrating Relation Reasoning for Content-aware Layout Generation with Multi-modal Large Language Models [7.288330685534444]
  We introduce Re, a novel method that leverages relation-CoT to generate more reasonable and coherent layouts.<n>Specifically, we enhance layout annotations by introducing explicit relation definitions, such as region, salient, and margin between elements.<n>We also introduce a layout prototype sampler, which defines layout prototype features across three dimensions and quantifies distinct layout styles.
  arXiv  Detail & Related papers  (2025-07-08T01:13:43Z)
• Illuminating Spaces: Deep Reinforcement Learning and Laser-Wall Partitioning for Architectural Layout Generation [0.0]
  This paper offers a procedural approach that intuitively mimics the process of human designers.<n>Effectively using RL for SLD requires an explorative space composing method to generate desirable design solutions.<n>We introduce "laser-wall", a novel space partitioning method that conceptualizes walls as emitters of imaginary light beams to partition spaces.
  arXiv  Detail & Related papers  (2025-02-06T09:35:24Z)
• LayoutVLM: Differentiable Optimization of 3D Layout via Vision-Language Models [57.92316645992816]
  Spatial reasoning is a fundamental aspect of human cognition, enabling intuitive understanding and manipulation of objects in three-dimensional space.<n>We introduce LayoutVLM, a framework and scene layout representation that exploits the semantic knowledge of Vision-Language Models (VLMs)<n>We demonstrate that fine-tuning VLMs with the proposed scene layout representation extracted from existing scene datasets can improve their reasoning performance.
  arXiv  Detail & Related papers  (2024-12-03T06:15:04Z)
• GLDesigner: Leveraging Multi-Modal LLMs as Designer for Enhanced Aesthetic Text Glyph Layouts [53.568057283934714]
  We propose a Vision-Language Model (VLM)-based framework that generates content-aware text logo layouts.<n>We introduce two model techniques that reduce the computational cost for processing multiple glyph images simultaneously.<n>To support instruction tuning of our model, we construct two extensive text logo datasets that are five times larger than existing public datasets.
  arXiv  Detail & Related papers  (2024-11-18T10:04:10Z)
• REVISION: Rendering Tools Enable Spatial Fidelity in Vision-Language Models [67.55362046790512]
  Vision-language models lack the ability to correctly reason over spatial relationships. We develop the REVISION framework which improves spatial fidelity in vision-language models. Our results and findings indicate that utilizing rendering-based frameworks is an effective approach for developing spatially-aware models.
  arXiv  Detail & Related papers  (2024-08-05T04:51:46Z)
• PosterLLaVa: Constructing a Unified Multi-modal Layout Generator with LLM [58.67882997399021]
  Our research introduces a unified framework for automated graphic layout generation.<n>Our data-driven method employs structured text (JSON format) and visual instruction tuning to generate layouts.<n>We develop an automated text-to-poster system that generates editable posters based on users' design intentions.
  arXiv  Detail & Related papers  (2024-06-05T03:05:52Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.