Accelerating Photonic Integrated Circuit Design: Traditional, ML and Quantum Methods

TL;DR

This paper reviews advancements in Photonic Integrated Circuit design, comparing traditional, machine learning, and quantum methods to improve scalability, efficiency, and address current design challenges.

Contribution

It provides a comprehensive comparison of traditional, ML, and quantum approaches, highlighting recent progress and future potential in PIC design workflows.

Findings

ML approaches improve design scalability and efficiency

Quantum algorithms show promise for PIC optimization

Traditional methods are slower and less scalable

Abstract

Photonic Integrated Circuits (PICs) provide superior speed, bandwidth, and energy efficiency, making them ideal for communication, sensing, and quantum computing applications. Despite their potential, PIC design workflows and integration lag behind those in electronics, calling for groundbreaking advancements. This review outlines the state of PIC design, comparing traditional simulation methods with machine learning approaches that enhance scalability and efficiency. It also explores the promise of quantum algorithms and quantum-inspired methods to address design challenges.