Novel optimization techniques such as nanophotonic inverse design are promising tools to significantly reduce the size of passive silicon photonic components while maintaining their functionality and performance. However, optimizing a single design from start to end is still time consuming when arbitrary geometric shapes are allowed within the design space. This project aims at speeding up the optimization process in this context, merging insights from physics, optimization and machine learning. Those improvements are expected to benefit the design process in general and allow generating large and useful datasets for subsequent application of pattern recognition methods in particular.
Project team
Dr. Yuri Grinberg
Dr. Yuri Grinberg is an associate research officer at the National Research Council of Canada and is an expert in applied and theoretical machine learning and reinforcement learning. His primary research interests are in advancing the machine learning state of the art to address problems in physics and engineering. In particular, he is interested in the development of appropriate AI methodologies for the design of efficient, small-footprint, easy-to-fabricate photonic components with significantly reduced human effort.
Find out more about Dr. Grinberg's research interests and his publications.
Dr. Dan-Xia Xu
Dr. Dan-Xia Xu is a principal research officer at the National Research Council of Canada, a Fellow of the Royal Society of Canada, a Fellow of the Optical Society (OSA), and an adjunct professor with Carleton University. Her research field focusses on silicon photonics for optical communications, photonic thermometry, molecular sensing and beyond. She has led the pioneering work in cladding stress engineering for polarization control of photonic components, and in high-sensitivity biosensor systems using Si wire spiral resonators. One particular recent interest is to transform the traditional photonic design and discovery cycle through the use of AI and machine learning (ML) methods. The ultimate goal of this work is to create AI/ML-enabled computational methods that augment designer's knowledge and intuition, accelerate the photonic design process, and enable the autonomous discovery of non-intuitive and high-performance designs.