This module will introduce essential concepts and basic building blocks of deep learning, including: neurons, layers, convolutions, activation functions, loss, backpropagation, optimizers; implementation and training of fully connected and convolutional neural networks; and overview of a few interpretability techniques (Visualizing filters, tSNE, GradCAM).
Day and Time: November 11, 3-5 p.m. US CT
This session is a part of the ALCF AI for Science Training Series.
About the Speaker:
Asad is a physics PhD Student in the gravity group at the National Center for Supercomputing Applications. He is interested in applying machine learning, and specifically deep learning techniques to accelerate discoveries in physics. Asad is also interested in the theoretical underpinnings of Computer Vision, NLP, Unsupervised Learning, and High Perfomance Parallel Computing. Before grad school, he studied Mathematics and Physics for Bachelors of Science at the University of Minnesota, Twin Cities.