Jonathan Byrne

Background

About Jonathan Byrne

I am a researcher with over 25 years of programming experience. I have written programs for machine learning, modelling natural systems, data-mining, game A.I. design optimization, data visualization, automatic mesh reconstruction, distributed search, parallelization on high end computers and many other things. There is an exhaustive list of my projects and many side projects on my website and the code is available on request. My PhD focused on design optimization using evolutionary algorithms. It works by initially generating random solutions, selecting the best ones and using those to create the next generation. It is an iterative approach that produces improved designs that were not explicitly specified by me and is capable of showing computational creativity. I applied this approach to optimize electricity pylons, bridges, and aircraft. My post-doc research focused on 3D printing and additive manufacturing. It was the natural progression for my research to automatically produce the resulting designs. I love the idea of digitizing an object, optimizing it and then printing the improved design. I was the research director of the 3D printing hub in UCD and had access to the first metal printer available for research in Ireland. I previously ran the advanced architecture team at Intel Movidius where we built applications that run on Movidius hardware. This includes compressed mapping formats (VOLA), deep learning optimisation through quantization and sparsity. We also supported an incubator to help companies build deep learning applications and move to Movidius technology I am now working on vision systems for the production facilities in Intel Leixlip.