K-TRAIL is a simulator-guided generative framework that uses diffusion-based layout generation combined with derivative-free ensemble Kalman guidance to design RF and electromagnetic circuits. It allows a black-box EM simulator to refine candidate layouts during generation, supporting synthesis from target S-parameter responses or direct RF performance constraints. Experiments on multi-layer RFIC structures demonstrate improved agreement with target responses and the ability to discover structurally distinct layouts that meet design requirements.
By Piyush Saha, Evan Newell, Hanna O'Leary, Arun Natarajan, Alireza Aghasi
arXiv:2602. 15648v2 Announce Type: replace Abstract: Inverse design problems are common in engineering and materials science.
By Jens U. Kreber, Christian Wei{\ss}enfels, Joerg Stueckler
The paper presents a conditional diffusion framework for the inverse design of dielectric resonator metasurfaces based on target angular scattering patterns. Trained on T‑matrix simulated geometry‑response pairs, the model learns a distribution of feasible geometries, allowing multiple candidate designs for the ill‑posed inverse problem. The best generated metasurface achieves a mean percentage error of 1.39%, outperforming CMA‑ES optimization and deterministic neural baselines, and can be inferred in about one minute after training.
By M. Tsukerman, K. Grotov, D. Vovchuk, P. Ginzburg
arXiv:2606. 29272v1 Announce Type: new Abstract: Technology computer-aided design (TCAD) semiconductor device simulation is fundamentally constrained by the high computational cost of iteratively solving coupled drift-diffusion equations.
By Yihan Zhang, Zhiteng Zhang, Kun Chen, Chen Wang
arXiv:2609.13819v1 Announce Type: new
Abstract: Solving inverse problems with differentiable physics simulators holds the potential to revolutionize scientific discovery and engineering design, as it...
By Xiang Chen, Huanhuan Xia
arXiv:2405.07944v1 Announce Type: cross
Abstract: Among the well-known methods to approximate derivatives of expectancies computed by Monte-Carlo simulations, averages of pathwise derivatives are oft...
By Max Aehle, Mih\'aly Nov\'ak, Vassil Vassilev, Nicolas R. Gauger, Lukas Heinrich, Michael Kagan, David Lange
The paper presents a method for training single‑step neural surrogates that can handle wave‑scattering inverse problems with tens of thousands of controllable variables. By dynamically generating training examples through gradient ascent and using a replay dataset with normalization, the authors achieve a surrogate that accurately models two‑dimensional wave scattering for up to 41,772 variables and can generalize to over 3 million variables without retraining. The surrogate demonstrates comparable or better performance than traditional FDTD simulations for large‑scale forward simulations and inverse design of photonic devices, achieving speedups up to 26.5×.
arXiv:2505. 06800v2 Announce Type: replace-cross Abstract: Diffusion-based generative models have renewed interest in stochastic differential equation methods for sampling from complex distributions.
By Jairon H. N. Batista, Fl\'avio B. Gon\c{c}alves, Yuri F. Saporito, Rodrigo S. Targino
arXiv:2605.11506v2 Announce Type: replace
Abstract: Score-based diffusion models achieve state-of-the-art performance for inverse problems, but their practical deployment is hindered by long inferenc...
By Julio Oscanoa, Irmak Sivgin, Cagan Alkan, Daniel Ennis, John Pauly, Mert Pilanci, Shreyas Vasanawala
The paper presents a method for training single‑step neural surrogates that can handle wave‑scattering problems with tens of thousands of controllable variables. By dynamically generating training examples that highlight surrogate errors and using a replay dataset with normalization, the authors achieve a surrogate that accurately simulates two‑dimensional wave scattering for up to 41,772 variables and generalizes to over 3 million variables without retraining. The surrogate is applied to forward simulations and inverse design of freeform beam splitters and gradient‑index lenses, achieving speedups up to 26.5× compared to traditional FDTD methods.
By Charles Dove, Laura Waller
The paper presents a surrogate‑assisted optimization framework for designing a seven‑parameter current‑excited electromagnetic coil, coupling a 2‑D axisymmetric FEM model with a Matern 5/2 Gaussian‑process surrogate. Sequential Bayesian optimization using expected improvement (EI) is compared with COBYLA and BOBYQA, showing that the ranking of methods depends on the FEM evaluation budget and that different methods excel at early progress, terminal response, or computational cost. A retrospective study indicates no clear advantage of EI over posterior‑mean ranking on this smooth response surface, and the results are specific to the axisymmetric benchmark used.
By Yucheng Liu
The paper introduces a physics‑informed graph attention network that directly operates on the tetrahedral mesh used in TCAD simulations of FinFET devices. By predicting electrostatic potential and quasi‑Fermi levels at every mesh node and training with both data loss and finite‑volume current‑continuity residuals, the surrogate retains the underlying carrier‑transport physics while achieving size generalization. Benchmarks against Sentaurus Device show sub‑volt RMSE for the drift‑diffusion fields and a per‑design throughput that is orders of magnitude faster, enabling rapid Pareto‑front exploration of large multi‑fin arrays that would otherwise be prohibitively slow to simulate.
By Leonid Popryho, Ayoub Sadeghi, Inna Partin-Vaisband