ALCF DeepHyper Automated Machine Learning Workshop

Argonne National Laboratory

July 15, 2022

The Argonne Leadership Computing Facility (ALCF) will host a hands-on training session on DeepHyper (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).

Organizers

Prasanna Balaprakash

Romain Egele

Romit Maulik

Bethany Lusch

Krishnan Raghavan

Sandeep Madireddy

Tanwi Mallick

Nesar Soorve Ramachandra

Kyle Felker

Anirudh Subramanyam

Akhil Pandey Akella

Joceran Gouneau

Sam Foreman

Event details

In this virtual workshop attendees will learn various capabilities of the DeepHyper software to automate the design and development of neural networks.

Workshop repo: github.com/deephyper/anl-22-summer-workshop
DeepHyper GitHub repo: github.com/deephyper/deephyper
DeepHyper Documentation: deephyper.readthedocs.io

All times are Central Standard Time (CDT)

Part 1

8:30 – 9:00: Checkin and setup
9:00 – 9:10: Welcome & Intro (Prasanna Balaprakash) [Slides][Video]
9:10 – 9:30: Hyperparameter search (Prasanna Balaprakash) [Slides][Video]
9:30 – 10:00: Hands-on (Romain Egele) [Video]
10:00 – 10:20: Neural architecture search (Romit Maulik) [Slides][Video]
10:20 – 10:50: Hands-on (Romit Maulik) [Video]

Break for 10 mins

Part 2

11:00 – 11:20: Ensembles & uncertainty quantification (Bethany Lusch and Krishnan Raghavan) [Slides][Video]
11:20 – 11:50: Hands-on (Bethany Lusch and Krishnan Raghavan) [Video]

Lunch Break and Q & A session for 40 mins

Part 3

12:30 – 12:50: Transfer learning in hyperparameter search (Sandeep Madireddy and Tanwi Mallick) [Slides][Video]
12:50 – 1:20: Hands-on (Sandeep Madireddy and Tanwi Mallick) [Video]
1:20 – 2:10: Running DeepHyper at Scale (No handson; only demo)[Video]
- NERSC-Perlmutter (Nesar Soorve Ramachandra) [Slides,Example]
- ALCF-ThetaGPU (Sam Foreman, Joceran Gouneau) [Slides,Example]
- OLCF-Summit platforms (Kyle Felker) [Slides,Example]

Break for 10 mins

Part 4 - Latest Research

2:20 – 2:40: Multiobjective hyperparameter search (Anirudh Subramanyam) [Slides] [Video]
2:40 – 3:10: Hands-on (Anirudh Subramanyam) [Video]
3:10 – 3:30: Graph neural architecture search for molecular property prediction (Akhil Pandey Akella) [Slides][Video]
3:30 – 4:00: Hands-on (Akhil Pandey Akella) [Video]