Photonic integrated circuits are facilitating the development of optical neural networks, which have the potential to be both faster and more energy efficient than their electronic counterparts since optical signals are especially well-suited for implementing matrix multiplications. However, accurate programming of photonic chips for optical matrix multiplication remains a difficult challenge. Here, we describe both simple analytical models and data-driven models for offline training of optical matrix multipliers. We train and evaluate the models using experimental data obtained from a fabricated chip featuring a Mach-Zehnder interferometer mesh implementing 3-by-3 matrix multiplication. The neural network-based models outperform the simple physics-based models in terms of prediction error. Furthermore, the neural network models are also able to predict the spectral variations in the matrix weights for up to 100 frequency channels covering the C-band. The use of neural network models for programming the chip for optical matrix multiplication yields increased performance on multiple machine learning tasks.

No abstract

<p>We quantify thermal crosstalk in a programmable photonic processor and present both analytical and data-driven models. We experimentally demonstrate model-based predictive crosstalk compensation for a microring resonator realized on a pre-calibrated chip, making it possible to tune the resonance wavelength with ±0.5 picometer accuracy.</p>

<p>We quantify thermal crosstalk in a programmable photonic processor and present both analytical and data-driven models. We experimentally demonstrate model-based predictive crosstalk compensation for a microring resonator realized on a pre-calibrated chip, making it possible to tune the resonance wavelength with ±0.5 picometer accuracy.</p>