ALCF DeepHyper Automated Machine Learning Workshop
The Argonne Leadership Computing Facility will host a hands-on training session on DeepHyper (https://github.com/deephyper/deephyper), a distributed automated machine learning (AutoML) software package for automating the design and development of deep neural networks for scientific and engineering applications.
DeepHyper seeks to bring a scientifically rigorous automated approach to neural network model development by reducing the manually intensive, cumbersome, ad hoc, trial-and-error efforts.
DeepHyper can run on a laptop, medium range clusters, and supercomputers with thousands of compute units (GPUs).
Event Details
In this virtual workshop attendees will learn various capabilities of the DeepHyper software to automate the design and development of neural networks.
DeepHyper GitHub repo: https://github.com/deephyper/deephyper
DeepHyper Documentation: https://deephyper.readthedocs.io/en/latest/
Agenda
Times |
Topics |
Speakers |
| 8:30-9:00am (CST) | Attendee Check-in | |
| 9:00-9:10am (CST) | Welcome & Introduction | Prasanna Balaprakash (Argonne) |
| 9:10-9:30am (CST) | Hyperparameter Search | Prasanna Balaprakash (Argonne) |
| 9:30-10:00am (CST) | Hands-on | Romain Égelé (Argonne) |
| 10:00-10:20am (CST) | Neural Architecture Search | Romit Maulik (Argonne) |
| 10:20-10:50am (CST) | Hands-on | Romit Maulik (Argonne) |
| 10:50-11:00am (CST) | Break | |
| 11:00am-11:20am (CST) | Ensembles & Uncertainty Quantification | Bethany Lusch and Krishnan Raghavan (Argonne) |
| 11:20-11:50am (CST) | Hands-on | Bethany Lusch and Krishnan Raghavan (Argonne) |
| 11:50-12:30pm (CST) | Lunch Break and Q & A Session for 40 mins | |
|
12:30-12:50pm (CST) |
Multi-Objective Hyperparameter Search | Anirudh Subramanyam (Argonne) |
| 12:50-1:20PM (CST) | Hands-on | Anirudh Subramanyam (Argonne) |
| 1:20-1:40pm (CST) | Transfer Learning in Hyperparameter Search | Sandeep Madireddy and Tanwi Mallic (Argonne) |
| 1:40-2:10pm (CST) | Hands-on | Sandeep Madireddy and Tanwi Mallick (Argonne) |
| 2:10-2:20pm (CST) | Break | |
| 2:20-2:40pm (CST) | Graph Neural Architecture Search for Molecular Property Prediction | Akhil Pandey Akella (Northern Illinois University) |
| 2:40-3:10pm (CST) | Hands-on | Akhil Pandey Akella (Northern Illinois University) |
| 3:10-4:00pm (CST) | Running DeepHyper at Scale on Perlmutter/NERSC, ThetaGPU/ALCF, Summit/OLCF Platforms | Nesar Soorve Ramachandra, Kyle Felker, and Sam Foreman (Argonne) |