Wave Interpolation Neural Operator: Interpolated Prediction of Electric Fields Across Untrained Wavelengths

2024-08-06

arXiv.org

AI Summary

Conventional electromagnetic surrogate solvers exhibit poor generalization across wavelengths, limiting their applicability in photonic design. Method: We propose the first wavelength-continuous generalizable electromagnetic field surrogate solver for photonic structure design, enabling zero-shot interpolation of electric field distributions at arbitrary unseen wavelengths using only discretely sampled training wavelengths. Our approach innovatively integrates the Fourier Neural Operator framework with a group-equivariant convolutional rearrangement operator and introduces a frequency-domain conditional encoding mechanism to explicitly model and efficiently generalize across the wavelength dimension. Contributions/Results: Experiments demonstrate a ~100× speedup over FDFD-based solvers, a 74% reduction in trainable parameters, and significant accuracy improvements: +80.5% on untrained wavelengths and +13.2% on trained wavelengths—substantially advancing both parameter efficiency and cross-wavelength generalization capability.

Technology Category

• Search and Optimization: Mixed Discrete/Continuous Search
• Computer Vision: Learning & Optimization for CV
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Sustainability and carbon-aware systems for Web, mobile, and WoT
• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines

Abstract

Designing photonic structures requires electromagnetic simulations, which often require high computational costs. Researchers have developed surrogate solvers for predicting electric fields to alleviate the computational issues. However, existing surrogate solvers are limited to performing inference at fixed simulation conditions and require retraining for different conditions. To address this, we propose Wave Interpolation Neural Operator (WINO), a novel surrogate solver enabling simulation condition interpolation across a continuous spectrum of broadband wavelengths. WINO introduces the Fourier Group Convolution Shuffling operator and a new conditioning method to efficiently predict electric fields from both trained and untrained wavelength data, achieving significant improvements in parameter efficiency and spectral interpolation performance. Our model demonstrates approximately 100 times faster performance than traditional finite-difference frequency-domain simulations. Moreover, compared to the state-of-the-art model, we achieve a 74% reduction in parameters and 80.5% improvements in prediction accuracy for untrained wavelengths, and 13.2% improvements for trained wavelengths.

Problem

• Develops a broadband electromagnetic surrogate solver
• Improves prediction accuracy for untrained wavelengths
• Reduces model parameters while enhancing generalization

Innovation

• Broadband surrogate solver for continuous wavelengths
• Wave-Informed Multiplicative Encoding reduces overfitting
• Fourier Group Convolutional Shuffling enhances accuracy