arXiv：cs.LG· Pawe{\l} Skier\'s, Ma{\l}gorzata Grzanka, Wojciech Masarczyk, Jan-Willem van de Meent, Kamil Deja·· 4 小时前AI 评分51
PaTh 论文提出递归潜空间推理方法提升扩散模型视觉推理能力
Think Before You Paint: Recursive Latent Reasoning for Diffusion Models
AI 导读
论文提出 Painter-Thinker（PaTh），用小型递归网络（Thinker）在每个去噪步内细化潜状态，并通过 ControlNet 适配器引导冻结的扩散模型（Painter），仅用重建损失训练，无需符号目标、求解器或验证器。
正文
Abstract:Diffusion models generate realistic images but often fail on visual reasoning tasks, such as filling in a Sudoku or drawing the path through a maze. When a discrete symbolic representation is available, recursive methods such as the Tiny Recursive Model (TRM) solve even hard instances of these puzzles. We ask how such reasoning can be carried over to pixels, where no symbolic representation is available. We propose Painter-Thinker (PaTh): a small recursive network (the Thinker) reasons over a grid of learned tokens that encode the noisy image and the conditioning, refines a latent state within every denoising step, and steers a frozen diffusion model (the Painter) through ControlNet adapters. The Thinker is trained with the standard reconstruction loss alone, without symbolic targets, a solver, or a verifier. PaTh solves 92.5% of hard MNIST Sudoku puzzles (prior best 75%) and 71.2% of extreme ones (prior best 4.1%), with 10M parameters against 82M for a standard diffusion model. It also improves on mazes, Queens, and CLEVR scenes with specified spatial relations, and its advantage grows with problem size. Diagnostic experiments show that PaTh recovers from injected mistakes that the diffusion model cannot repair, especially when many cells are wrong. Together, these results show that reasoning mechanisms developed for symbolic data can be integrated into pixel-space diffusion without symbolic supervision, opening a path toward generating data under increasingly complex constraints.
| Subjects: | Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.09876 [cs.AI] |
| (or arXiv:2610.09876v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.09876 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Paweł Skierś [view email]
[v1]
Wed, 7 Oct 2026 11:36:56 UTC (3,817 KB)
来源：arXiv：cs.LG · arxiv.org