Brian
McFadden
ML Solutions Engineer
Edge Impulse
Whether it’s been programming embedded control systems, assembling and testing innovative biotech, proving out a novel microscopy illumination concept, working on autonomous underwater vehicles, collaborating with international manufacturers, or launching new products, Brian’s career has been about bringing technically challenging ideas into reality — and that’s what he loves about engineering. In addition to working at Edge Impulse, Brian is currently pursuing a MASc in machine learning as part of the Intelligent Sensing, Diagnostics, and Prognostics lab at the University of British Columbia.
16 April 2024 11:45 - 12:15
Help! I don’t have enough data...how to rapidly validate an idea when starting from scratch
Imagine this scenario: you have an idea for a computer vision based defect detection system that is going to save your company significant money. Alright! Let’s train a model and try it out... Shoot, you don’t have any data, and data collection can be costly and time consuming. So what do you do? In this talk Brian will address this common problem by discussing how to quickly test concepts through showing the benefits of visual anomaly detection with FOMO-AD and covering the generation of synthetic data.