Paper
From "What" to "How": Constrained Reasoning for Autoregressive Image Generation
Authors
Ruxue Yan, Xubo Liu, Wenya Guo, Zhengkun Zhang, Ying Zhang, Xiaojie Yuan
Abstract
Autoregressive image generation has seen recent improvements with the introduction of chain-of-thought and reinforcement learning. However, current methods merely specify "What" details to depict by rewriting the input prompt, yet fundamentally fail to reason about "How" to structure the overall image. This inherent limitation gives rise to persistent issues, such as spatial ambiguity directly causing unrealistic object overlaps. To bridge this gap, we propose CoR-Painter, a novel framework that pioneers a "How-to-What" paradigm by introducing Constrained Reasoning to guide the autoregressive generation. Specifically, it first deduces "How to draw" by deriving a set of visual constraints from the input prompt, which explicitly govern spatial relationships, key attributes, and compositional rules. These constraints steer the subsequent generation of a detailed description "What to draw", providing a structurally sound and coherent basis for accurate visual synthesis. Additionally, we introduce a Dual-Objective GRPO strategy that specifically optimizes the textual constrained reasoning and visual projection processes to ensure the coherence and quality of the entire generation pipeline. Extensive experiments on T2I-CompBench, GenEval, and WISE demonstrate that our method achieves state-of-the-art performance, with significant improvements in spatial metrics (e.g., +5.41% on T2I-CompBench).
Metadata
Related papers
Vibe Coding XR: Accelerating AI + XR Prototyping with XR Blocks and Gemini
Ruofei Du, Benjamin Hersh, David Li, Nels Numan, Xun Qian, Yanhe Chen, Zhongy... • 2026-03-25
Comparing Developer and LLM Biases in Code Evaluation
Aditya Mittal, Ryan Shar, Zichu Wu, Shyam Agarwal, Tongshuang Wu, Chris Donah... • 2026-03-25
The Stochastic Gap: A Markovian Framework for Pre-Deployment Reliability and Oversight-Cost Auditing in Agentic Artificial Intelligence
Biplab Pal, Santanu Bhattacharya • 2026-03-25
Retrieval Improvements Do Not Guarantee Better Answers: A Study of RAG for AI Policy QA
Saahil Mathur, Ryan David Rittner, Vedant Ajit Thakur, Daniel Stuart Schiff, ... • 2026-03-25
MARCH: Multi-Agent Reinforced Self-Check for LLM Hallucination
Zhuo Li, Yupeng Zhang, Pengyu Cheng, Jiajun Song, Mengyu Zhou, Hao Li, Shujie... • 2026-03-25
Raw Data (Debug)
{
"raw_xml": "<entry>\n <id>http://arxiv.org/abs/2603.02712v1</id>\n <title>From \"What\" to \"How\": Constrained Reasoning for Autoregressive Image Generation</title>\n <updated>2026-03-03T08:03:18Z</updated>\n <link href='https://arxiv.org/abs/2603.02712v1' rel='alternate' type='text/html'/>\n <link href='https://arxiv.org/pdf/2603.02712v1' rel='related' title='pdf' type='application/pdf'/>\n <summary>Autoregressive image generation has seen recent improvements with the introduction of chain-of-thought and reinforcement learning. However, current methods merely specify \"What\" details to depict by rewriting the input prompt, yet fundamentally fail to reason about \"How\" to structure the overall image. This inherent limitation gives rise to persistent issues, such as spatial ambiguity directly causing unrealistic object overlaps. To bridge this gap, we propose CoR-Painter, a novel framework that pioneers a \"How-to-What\" paradigm by introducing Constrained Reasoning to guide the autoregressive generation. Specifically, it first deduces \"How to draw\" by deriving a set of visual constraints from the input prompt, which explicitly govern spatial relationships, key attributes, and compositional rules. These constraints steer the subsequent generation of a detailed description \"What to draw\", providing a structurally sound and coherent basis for accurate visual synthesis. Additionally, we introduce a Dual-Objective GRPO strategy that specifically optimizes the textual constrained reasoning and visual projection processes to ensure the coherence and quality of the entire generation pipeline. Extensive experiments on T2I-CompBench, GenEval, and WISE demonstrate that our method achieves state-of-the-art performance, with significant improvements in spatial metrics (e.g., +5.41% on T2I-CompBench).</summary>\n <category scheme='http://arxiv.org/schemas/atom' term='cs.CV'/>\n <category scheme='http://arxiv.org/schemas/atom' term='cs.MM'/>\n <category scheme='http://arxiv.org/schemas/atom' term='eess.IV'/>\n <published>2026-03-03T08:03:18Z</published>\n <arxiv:primary_category term='cs.CV'/>\n <author>\n <name>Ruxue Yan</name>\n </author>\n <author>\n <name>Xubo Liu</name>\n </author>\n <author>\n <name>Wenya Guo</name>\n </author>\n <author>\n <name>Zhengkun Zhang</name>\n </author>\n <author>\n <name>Ying Zhang</name>\n </author>\n <author>\n <name>Xiaojie Yuan</name>\n </author>\n </entry>"
}