Welcome to my website & blog (link at the top)! My name is Matthew Kowal. I am an MSc in the Ryerson Vision Lab in Toronto, co-supervised by Dr. Neil Bruce and Dr. Kosta Derpanis. My research focuses on how we can better understand the world around us using neural networks and unlabelled data. In particular, I am currently working on action recognition with multi-modal and unsupervised data as well as semantic segmentation using weak supervision. I am also interested in how information theory can be used to better formulate traditional machine learning problems. I write blog posts every so often describing the things I am using in my research.


  • September, 2020 - Starting my PhD under the supervision of Dr. Kosta Derpanis @ Ryerson Vision Lab!

  • August, 2020 - Paper accepted to BMVC 2020 as an Oral presentation! Category Specific Feature Binding for Semantic Segmentation and Adversarial Robustness. Amirul Islam, Matthew Kowal, Kosta Derpanis, and Neil Bruce. Paper

  • May, 2020 - Awarded Ontario Graduate Scholarship (OGS)! (Accepted)

  • March, 2020 - Started role as AI Scientist in Residence @ NextAI, Toronto, Canada; technical and R&D consultant for multiple startups.

  • February, 2020 - Presented a poster titled ‘An Analysis of Structural Variant Callers’ at the Cancer Big Data and AI conference held in Toronto, Canada. Presented alongside Denis Keimakh (MSc @ The University of Toronto).

  • September, 2018 - Joined the Ryerson Vision Lab as an MSc, co-supervised by Dr. Neil Bruce and Dr. Kosta Derpanis.