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

arXiv: https://arxiv.org/abs/2602.13912 · HTML: https://mlsanity.com/paper/2602.13912/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Spatial and embodied reasoning, one of 17 directions in LLM reasoning.
- Builds on 22 of its 27 references that are in the map.
- First posted on arXiv 14 Feb 2026.

## Placed in

- LLM reasoning / [Spatial and embodied reasoning](https://mlsanity.com/lr/spatial-and-embodied-reasoning/)

## Builds on (most cited first, 22 of 22)

- [SpatialVLM: Endowing Vision-Language Models with Spatial Reasoning Capabilities](https://mlsanity.com/paper/2401.12168/) (2024)
- [Image Segmentation Using Text and Image Prompts](https://mlsanity.com/paper/2112.10003/) (2021)
- [Spatial-MLLM: Boosting MLLM Capabilities in Visual-based Spatial Intelligence](https://mlsanity.com/paper/2505.23747/) (2025)
- [From LLM Reasoning to Autonomous AI Agents: A Comprehensive Review](https://mlsanity.com/paper/2504.19678/) (2025)
- [ShapeLLM: Universal 3D Object Understanding for Embodied Interaction](https://mlsanity.com/paper/2402.17766/) (2024)
- [LayoutGAN: Generating Graphic Layouts with Wireframe Discriminators](https://mlsanity.com/paper/1901.06767/) (2019)
- [LayoutVAE: Stochastic Scene Layout Generation From a Label Set](https://mlsanity.com/paper/1907.10719/) (2019)
- [PosterLayout: A New Benchmark and Approach for Content-aware Visual-Textual Presentation Layout](https://mlsanity.com/paper/2303.15937/) (2023)
- [LayoutDM: Transformer-based Diffusion Model for Layout Generation](https://mlsanity.com/paper/2305.02567/) (2023)
- [SpatialPIN: Enhancing Spatial Reasoning Capabilities of Vision-Language Models through Prompting and Interacting 3D Priors](https://mlsanity.com/paper/2403.13438/) (2024)
- [Learning to Localize Objects Improves Spatial Reasoning in Visual-LLMs](https://mlsanity.com/paper/2404.07449/) (2024)
- [How to Enable LLM with 3D Capacity? A Survey of Spatial Reasoning in LLM](https://mlsanity.com/paper/2504.05786/) (2025)
- [RATT: A Thought Structure for Coherent and Correct LLM Reasoning](https://mlsanity.com/paper/2406.02746/) (2024)
- [Enhancing Decision-Making for LLM Agents via Step-Level Q-Value Models](https://mlsanity.com/paper/2409.09345/) (2024)
- [Large Language Models are Good Relational Learners](https://mlsanity.com/paper/2506.05725/) (2025)
- [LayoutNUWA: Revealing the Hidden Layout Expertise of Large Language Models](https://mlsanity.com/paper/2309.09506/) (2023)
- [Symbol-LLM: Leverage Language Models for Symbolic System in Visual Human Activity Reasoning](https://mlsanity.com/paper/2311.17365/) (2023)
- [PosterO: Structuring Layout Trees to Enable Language Models in Generalized Content-Aware Layout Generation](https://mlsanity.com/paper/2505.07843/) (2025)
- [BannerAgency: Advertising Banner Design with Multimodal LLM Agents](https://mlsanity.com/paper/2503.11060/) (2025)
- [CAL-RAG: Retrieval-Augmented Multi-Agent Generation for Content-Aware Layout Design](https://mlsanity.com/paper/2506.21934/) (2025)
- [Analyzing Multimodal Interaction Strategies for LLM-Assisted Manipulation of 3D Scenes](https://mlsanity.com/paper/2410.22177/) (2024)
- [LayoutRectifier: An Optimization-based Post-processing for Graphic Design Layout Generation](https://mlsanity.com/paper/2508.11177/) (2025)

## Built on by (most cited first, 1 of 1)

- [GDB-Reward: From Evaluation Metrics to Training Rewards for Graphic Design](https://mlsanity.com/paper/2609.02813/) (2026)

## Abstract (arXiv)

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.