Patrick Beukema

Hi! My name is Patrick Beukema. I am scientist and engineer passionate about applying machine learning and artificial intelligence to global environmental challenges. I am also interested in increasing the adoption of software engineering and data science skills in academia in order to improve scientific reproducibility. I currently work on deep neural network architectures for computer vision and natural language processing, often in combination, and serving those models scale.
I hold a Ph.D in neuroscience from the Center for Neuroscience at the University of Pittsburgh and the Center for the Neural Basis of Cognition at Carnegie Mellon University. My research focused on the evolution of biological neural networks during learning. Before that, I studied recurrent networks and causal inference at Carnegie Mellon University (M.S. Logic). And before that, I spent several great years at McGill, and learned the art of mixing Montreal winters and fixed gears.
Oct 6, 2020 | Talk: “Scalable document AI with CV, OCR, and NLP” |
Dec 16, 2018 | Medium post: “Modernizing Academic Data Science” |
Nov 18, 2018 | OpenNeuro datasets from dissertation [dataset1, dataset2]. |
Aug 24, 2018 | Workshop on reproducible data science, NeuroHackademy workshop materials. |
Jan 29, 2018 | “Plasticity of neuronal representations” on bioRxiv [data] [ipynb] [manuscript] |
Jan 17, 2018 | “Parallel motor learning networks”, Cosyne 2018. |
Nov 15, 2017 | “Algorithms supporting skill consolidation” Current Opinion |
Oct 30, 2017 | Talk “Statistical Abuse in Biomedical Research” [slides] |