Schedule June 2-3

9:00 AM

Canada/Eastern

Welcoming participants

9:30 AM

Canada/Eastern

Opening Remarks

9:30 a.m. Sophie Larochelle - Full Professor, Director, COPL 9:40 a.m. Paul Charrette - Full Professor, Director, 3IT

9:45 AM

Canada/Eastern

Shedding light on AI and vice versa

Jean-Francois Lalonde - Professor, Department of Computer and Electrical Engineering, Université Laval Abstract Artificial intelligence (AI) is playing an increasingly important role in many fields of science today, as a tool for analysis, modeling and discovery. In this presentation, I will offer a rigorous yet accessible insight into what AI really is, its main operating principles, and the types of approaches most commonly used in the research world. After shedding light on AI, we'll go the other way and explore how AI can be used to understand light. In particular, I will present work that estimates the properties of light sources in an image, facilitates light capture, and simulates the appearance of our visual world. Biography Jean-François Lalonde, Ph.D., is a full professor in the Department of Electrical and Computer Engineering at Université Laval and Associate Scientific Director of the Institute Intelligence and Data (IID). Before starting his career in 2013, he was a postdoctoral researcher at Disney Research. He received his PhD in robotics from Carnegie Mellon University (Pittsburgh, USA) in 2011. His research interests lie at the intersection of computer vision, computer graphics, and machine learning. He is particularly interested in how physical models can be combined with data-driven methods to better interpret, visualize, and synthesize the visual world. To this end, his group has captured and published the largest public datasets of indoor and outdoor high dynamic range panoramic images, freely available for research. His work has been disseminated through 100 scientific articles that have accumulated more than 7,500 citations. He is actively involved in transferring his research results to industry, as demonstrated by his 12 patents, his collaborations with companies such as Adobe and Meta and his role as scientific consultant for several high-tech companies.

10:45 AM

Canada/Eastern

Photonics for AI, AI for Photonics: Topology-Optimized Design and AI-Corrected Fabrication for Next-Gen Computing

Odile Liboiron-Ladouceur - Professor at McGill University Summary This talk explores how silicon photonic interconnects can be scaled using mode-division multiplexing (MDM), leveraging the high-index contrast of silicon-on-insulator (SOI) technology platforms. Through topology-optimized design methodologies, we demonstrate compact and broadband mode converters, efficient mode exchangers, and ultra-short lens-assisted tapers. In parallel, we present a machine learning framework that compensates for process-induced variability, significantly improving fabrication fidelity and device performance. These results highlight a bidirectional synergy between photonics and artificial intelligence: while photonics enables scalable hardware for AI, AI in turn enhances the design and manufacturability of photonic systems. This co-evolution paves the way for robust, compact, and energy-efficient optical solutions in next-generation computing architectures. Biography Odile Liboiron-Ladouceur is Full Professor in the Department of Electrical and Computer Engineering at McGill University and associate member of COPL. Her research focuses on photonic integrated circuits, AI-driven photonic design, and optical interconnects for high-performance computing. She has authored over 250 publications and holds seven granted U.S. patents, along with multiple international (PCT) and provisional applications. Her innovations have advanced scalable silicon photonics and energy-efficient optical networks. A co-founder of a photonics startup, she was named to the 2025 Photonics100 for her leadership in the field.

11:30 AM

Canada/Eastern

Practical application of AI: the case of optical design

Simon Thibault - Full Professor, Director of the Department of Physics, Engineering Physics and Optics and Optical Design Researcher, Université Laval Summary The advent of artificial intelligence (AI) is increasingly disrupting our habits. Is it possible to take advantage of this, specifically in optical design? We propose to explore these questions through concrete examples. The first example is LensNet, a starting point generator for optical design. This world-renowned optical design tool offers a positive first experience, even for teaching purposes. Then we'll look at how LensNet can be specialized to become a formidable optical designer. Finally, we'll look at how we've imagined using generative AI in optical design. Biography Simon Thibault is Full Professor and Head of the Department of Physics, Engineering Physics and Optics at Université Laval. A Fellow of SPIE and OPTICA, and recipient of the Summa recherche 2024 award (FSG-ULaval), he holds the NSERC Industrial Research Chair in Optical Design (http://lrio.copl.ulaval.ca/) and sits on several international conference program committees at OPTICA, SPIE and IEEE. He is the author of over 350 scientific papers and holds some 30 patents. He is associate editor of Optics Express. His research interests range from the use of metasurface and AI in optical design to the development of optical instruments for extreme environments.

12:15 PM

Canada/Eastern

Lunch Break

1:15 PM

Canada/Eastern

Introducing new members

Sophie Larochelle - Full Professor, Director, COPL

1:18 PM

Canada/Eastern

In memory of Michel Piché

Michel Olivier Professor, Department of Physics, Cégep Garneau

1:30 PM

Canada/Eastern

Tribute Award 2024 - Ming Li

Ming Li - Full professor, Chinese Academy of Sciences, Beijing, China

1:40 PM

Canada/Eastern

Charge carrier dynamics in quantum dots: from modeling to applications

Gabriella Tessitore - Professor, Department of Chemistry, Université Laval Summary (Free translation from French*)Modeling light-matter interactions in nanomaterials is an indispensable tool for understanding the complexity of charge carrier dynamics and enabling their exploitation. Machine learning algorithms are proving particularly useful for optimizing the synthesis of nanomaterials based on their optical properties. However, these algorithms often do not incorporate a detailed mathematical model of charge carrier dynamics, as their implementation would be impractical due to high processing times. Consequently, less information can be extracted about the mechanism itself. The modeling of luminescence dynamics by systems of differential equations is widely used and offers the possibility of considering the ionic distribution of an active species within a crystal lattice and its symmetry. Predicting the effect of the slightest change in the ionic distribution on the population of energy levels involved requires a higher level of sophistication, which would be accessible by coupling rate equation models taking into account crystal lattice occupancy and machine learning algorithms. However, as mentioned above, these models need to be considerably simplified to enable any resolution. A possible solution to this problem would be to find an approximate resolution of the rate equation models resulting in a known error, in order to integrate it into the final analysis of the mechanism. Ideally, an algebraic expression would be preferable to avoid excessive computational burden. In this quest for a suitable resolution, a compromise between accuracy and computational burden will be presented. The prospective application of such a model to the nanomaterials studied in our group will be clarified, in order to examine the potential implications of the use of quantum dots and their charge carrier dynamics in the biological, environmental and radiation safety fields.