Presenter: Asad Khan, University of Illinois

Tuesday, April 27, 2021

Slides: https://github.com/khanx169/NGA_NFI_webinar

Video: https://uofi.app.box.com/s/alwsvboa61wdzmm8utojy9zgxr0bsya7  (Part 1)

Abstract

This tutorial provides a hands-on introduction to basic deep learning concepts. Attendees will learn in a practical, set-by-step manner the key ideas that led to the development and use of deep neural networks, and will acquire a basic understating of the training and use of deep neural networks using state-of-the-art GPUs in the context of image classification. The tutorial will equip the participants with sufficient knowledge to quickly build, train/fine-tune, and test state-of-the-art machine learning models on a variety of computer vision tasks.

Prerequisites: Python, basic Linux. No prior HPC or Machine Learning experience is needed.

Preparations in Advance

The hands-on activities will be conducted using the HAL system at the University of Illinois. Participants will need to register to use this system by completing this form in advance of the session.

Participants will also need to load DUO onto a smart device for 2 factor authentication for logging into the HAL system. Participants are asked to follow the instructions for loading DUO posted at this site DUO in advance of the session.

Biography

Asad Khan is a physics Ph.D. student at the University of Illinois Urbana-Champaign and is a member of the Center for Artificial Intelligence Innovation at the National Center for Supercomputing Applications. His research focus lies at the intersection of Gravitational Waves, Deep Learning, and High-Performance Computing. Previously he obtained his Bachelors of Science in Physics and Mathematics at the University of Minnesota. He is broadly interested in harnessing Artificial Intelligence at scale to accelerate discoveries in science.