arXiv：cs.LG（机器学习，全量分类）· Jinhao Liang, Jacob K. Christopher, Michael Frei, Tommaso Dreossi, Nando Fioretto·· 5 小时前AI 评分33
Simulator-Refined Diffusion：用模拟器精炼扩散模型做射频逆向设计
Simulator-Refined Diffusion for Radio-Frequency Inverse Design
AI 导读
研究者提出 Simulator-Refined Diffusion（SRD），在扩散采样过程中结合低保真可微代理模型与高保真不可微全波电磁模拟器，用代理梯度提出扰动方向、再由模拟器沿该方向搜索有效设计更新。在 PCB 布局生成任务上，该方法持续优于现有 SOTA，分布内目标的仿真 S-parameters 与目标规格接近度提升最多 21.2%，分布外目标提升最多 19.8%。
正文
Abstract:Diffusion models have shown potential in inverse design of printed circuit boards (PCBs), enabling the generation of layouts conditioned on target S-parameters. Despite this promise, applying diffusion models to PCB layout generation remains challenging due to their difficulty in meeting the quantitative electromagnetic specifications. A common approach is gradient-based guidance, which biases the diffusion sampling process with the gradient of an objective used for evaluation. However, full-wave electromagnetic simulators are accurate but expensive and typically non-differentiable, whereas differentiable surrogates are informative but not always reliable. To address these limitations, this paper proposes Simulator-Refined Diffusion (SRD), a novel combination of a low-fidelity differentiable surrogate and a high-fidelity non-differentiable simulator within the diffusion sampling process. Unlike standard zeroth-order optimization, which requires a great number of random perturbations, our approach uses the surrogate's gradient to propose the perturbation direction while the simulator then searches based on this direction to identify an effective design update. Experimental results across different settings show that this method consistently outperforms current state-of-the-art methods, producing layouts whose simulated S-parameters match the target specifications up to 21.2% closer for in-distribution targets and up to 19.8% for out-of-distribution targets.
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2609.38363 [cs.LG] |
| (or arXiv:2609.38363v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2609.38363 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Tommaso Dreossi [view email]
[v1]
Tue, 29 Sep 2026 18:26:40 UTC (3,069 KB)
来源：arXiv：cs.LG（机器学习，全量分类） · arxiv.org