Skip To Content

Optica Advanced Photonics Congress


Demystifying: Machine Learning

Sunday, 28 July: 09:00 – 17:00

Machine learning has been successfully applied in many use cases ranging from optical devices to optical networking. Even though there are numerous educational resources available, most do not apply the techniques to problems relevant to the optical field. The gap between theory and practice makes it challenging for early-career engineers and scientists to start applying machine learning for their problems.

Demystifying: Machine Learning – presented by Optica – aims to bridge the gap between theory and practice of applying machine learning to problems relevant to the optical field in an interactive and hands-on way.


The event will include an introduction to machine learning, covering the basic knowledge of supervised, semi-supervised and unsupervised learning. Then, we will move to a discussion of convolutional neural networks (CNNs) and the concepts that are used to build them. These concepts will be accompanied by practical implementations of the algorithms presented.

Examples will be provided to the attendees in a way that will allow them to run the code in their own workspace. Experimentation on the provided code guided by the instructors will allow the attendees to understand how different techniques are combined to solve a problem. We will discuss modifications that will allow the machine learning techniques studied to present better results and/or be applied to different use cases. Instructors will also share some pitfalls and mistakes made when applying machine learning, and how they overcame these.

The example code will be written in Python, and the attendees will run their code using Google Collab.



28 July

Part One:
Introduction to Machine Learning
Introduction to machine learning 09:00 - 09:45
Applying supervised learning for performance prediction and estimation 10:00 - 10:45
Applying semi-supervised and unsupervised learning for anomaly detection 11:00 - 11:45
Part Two:
Applications of Machine Learning
Introduction to convolutional neural networks (CNNs) 14:00 - 14:45
Applying a CNN for the identification of Young’s slit separation directly from a diffraction pattern. 15:00 - 17:00



Early-career engineers and scientists from academia or industry with limited or no prior experience with machine learning


There are no strict prerequisites for the attendees. Attendees need to bring their own laptop, and use their personal Google account to access the hands-on part. No specific software is necessary.


Ben Mills

University of Southampton

About the Speaker

Ben Mills is a Principal Research Fellow at the Optoelectronics Research Centre at the University of Southampton, UK. He leads a group at the interface of deep learning and laser materials processing, with a focus on the application of real time generative and convolutional neural networks for experimental optimization and control. He has around 150 publications, 20 invited talks and GBP 11 million in funding in the area of “smart lasers.”

Carlos Natalino

Chalmers University of Technology

About the Speaker

Carlos Natalino is a researcher with the Chalmers University of Technology, Gothenburg, Sweden. Over the past 6 years he has been applying machine learning techniques to solve various optical networking and security problems. He has co-authored more than 30 papers presenting machine learning solutions to optical network problems. He is the author of the Optical RL-Gym, an open-source toolkit for applying reinforcement learning to optical networking problems. He is a member of Optica and IEEE.


About the Optica Demystifying Series

The Optica Demystifying series offers introductions to the topics transforming science and engineering today. The objective is to lower the barriers to entry and get you started in the most exciting tools, techniques and areas of research in optics and photonics.  Whether you are just starting your career or well along your path, if you are looking for new opportunities, the leading researchers and great teachers in the Optica Demystifying series will help you take that important first step. Look for Demystifying offerings at your next Optica Event.

Image for keeping the session alive