Active learning for photonics

2026-01-22

AI Summary

This work proposes an active learning framework based on an analytically tractable last-layer Bayesian neural network (LL-BNN) to efficiently predict photonic bandgaps while reducing the computational cost of full-wave simulations. For the first time in photonics, the analytical LL-BNN is employed to guide uncertainty-driven sample selection, prioritizing the acquisition of structurally informative data for simulation-based training. The resulting uncertainty estimates exhibit strong correlation with actual prediction errors, substantially enhancing data efficiency. In the task of regressing bandgaps of two-dimensional bichromatic photonic crystals, the proposed approach achieves comparable prediction accuracy to random sampling while reducing the required training data by a factor of 2.6.

Technology Category

• Machine Learning: Active Learning
• Search and Optimization: Sampling/Simulation-based Search
• Reasoning under Uncertainty: Stochastic Optimization

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Web Mining and Content Analysis: Web data generation and simulation
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Active learning for photonic crystals explores the integration of analytic approximate Bayesian last layer neural networks (LL-BNNs) with uncertainty-driven sample selection to accelerate photonic band gap prediction. We employ an analytic LL-BNN formulation, corresponding to the infinite Monte Carlo sample limit, to obtain uncertainty estimates that are strongly correlated with the true predictive error on unlabeled candidate structures. These uncertainty scores drive an active learning strategy that prioritizes the most informative simulations during training. Applied to the task of predicting band gap sizes in two-dimensional, two-tone photonic crystals, our approach achieves up to a 2.6x reduction in required training data compared to a random sampling baseline while maintaining predictive accuracy. The efficiency gains arise from concentrating computational resources on high uncertainty regions of the design space rather than sampling uniformly. Given the substantial cost of full band structure simulations, especially in three dimensions, this data efficiency enables rapid and scalable surrogate modeling. Our results suggest that analytic LL-BNN based active learning can substantially accelerate topological optimization and inverse design workflows for photonic crystals, and more broadly, offers a general framework for data efficient regression across scientific machine learning domains.

Problem

• active learning
• photonic crystals
• band gap prediction
• data efficiency
• surrogate modeling

Innovation

• active learning
• analytic LL-BNN
• uncertainty quantification
• photonic crystals
• data-efficient surrogate modeling