Deep Learning Method for Quantum Efficiency Reconstruction

Abstract

We suggest a new scheme for measuring the quantum efficiency of camera sensors based on the reflection from a variable width Fabry-Perot resonator and a deep learning algorithm, outperforming standart reconstruction methods.

Predicting Broadband Resonator-Waveguide Coupling for Microresonator Frequency Combs through Fully Connected and Recurrent Neural Networks and Attention Mechanism

“…The field of photonics also has seen a surge of interest in the use of artificial intelligence, particularly neural networks, to enhance the understanding and application of light–matter interactions. In particular, neural networks have already shown promise in a variety of photonics applications such as inverse photonic design, − material and device characterization, ,,− optical sensing, image processing and classification, and optical communication . In inverse photonic design, − the goal is to design optical components or devices with specific desired properties such as the desired transmission spectrum, scattering properties, bandwidth, or quantum efficiency.…”

“…For the forward problems in photonics, such as predicting the reflectance spectrum of nanoparticles, optical properties of photonic crystals, and photonic crystal fibers, or device performance metrics from device parameters, − fully connected and recurrent neural networks (FCNNs and RNNs, respectively) are two of the most common deep-learning architectures. Due to the relative simplicity of forward problems compared to the inverse problems, neural networks can make predictions with very high accuracy, typically 98% or even higher. , This high accuracy is beneficial for several types of problems in photonic design.…”

“…If high accuracy is obtained, then this procedure can be used to reduce the computation time typically spent during device design and optimization studies. Considering the proven success of deep learning in forward applications, − high accuracy can be obtained with different types of neural networks such as FCNNs and RNNs. Noting that FCNNs require a fixed input size by design, if one wants to work with varying size inputs such as Q c values calculated at varying frequency values and ranges, then RNNs would be the more appropriate choice.…”

Attention mechanisms for broadband feature prediction for electromagnetic and photonic applications

“…We have witnessed a surge of interest in the use of artificial intelligence, specifically neural networks, to enhance the understanding and application of light-matter interactions. [1][2][3][4][5][6][7][8] Neural networks have shown promise in various photonics applications, including inverse photonic design, material and device characterization, optical sensing, image processing and classification, and optical communication.…”

“…For forward problems in photonics, such as predicting optical properties or device performance metrics, fully connected and recurrent neural networks (FCNNs and RNNs, respectively) are commonly used due to their high accuracy. [1][2][3][4][5][6][7][8] Neural networks significantly reduce computation time compared to traditional simulation-based approaches, especially when dealing with weakly non-linear problems. Our focus in this work is on the behavior of coupling quality factor (Q c ) in ring resonators, which are critical elements in optics and photonics.…”