Arthur Boschet

Welcome to my website! I am an Applied AI Research Scientist at Optina Diagnostics, specializing in harnessing the latest advancements in deep learning and computer vision. My work focuses on developing innovative tools for optometrists, ophthalmologists, and the diagnosis of Alzheimer’s disease. Previously, during my Master’s program at Mila, I completed a research internship at Optina Diagnostics, mentored by Jean-Sébastien Grondin and Adam Ibrahim. My primary project involved the generation of high-resolution hyperspectral retinal images using diffusion models and VQ-GANs, enhancing a self-supervised learning pipeline.

I recently completed a research internship at NeuroPoly, guided by Julien Cohen-Adad, where I developed segmentation models for axons and myelin in nervous system histology images. During this internship, I also co-authored the paper Unpaired Modality Translation for Pseudo Labeling of Histology Images, presented at the MICCAI 2024 workshop on Deep Generative Models. This work introduces an adversarial diffusion pipeline for modality translation, specifically designed to address data scarcity by generating pseudo labels for medical imaging modalities lacking labeled data.

Prior to these experiences, I earned a Bachelor’s degree in Bioengineering with a minor in Computer Science from McGill University. During my undergraduate studies, I engaged in various projects, including the development of motor imagery EEG-based systems, analyzing sentiment in Instagram posts related to Covid, and designing an intelligent surgical drain system.