2.5b Update: The Application of Machine Learning and Artificial Intelligence on HPC Systems for Defence Research
Tracks
Track 2
| Wednesday, November 11, 2020 |
| 1:30 PM - 2:25 PM |
Speaker
Prof John Taylor
Chief Computational Scientist
DST
.
Abstract
I will explore the scalable applications of artificial intelligence on massive data sets. First, I will cover how we optimized and developed highly parallelized implementations of DL algorithms and tested them on HPC GPU clusters. Then I will demonstrate how to develop models that can run over large high-resolution datasets, identifying the spatial and temporal relationships between physical parameters in global-scale high-resolution numerical weather prediction models. I will explore the implications of this research for constructing large scale Ml/AI workflows for Defence research.
I will explore the critical role of HPC in supporting Defence AI/ML research. I will provide an update on the development of our new HPC capability for Defence Research and National Security to be deployed at DST. A key advantage of HPC technology is that it will dramatically lift researcher productivity and create new research opportunities. Research that would take years on a desktop environment can be done in hours or minutes on a modern HPC system. A modern HPC system supports traditional modelling and simulation tasks, machine learning and AI, and data analytics at massive scale and I will provide defence relevant examples of all these use cases. In conclusion, a modern HPC system will help to ensure that we maintain our competitive advantage in what is a highly competitive environment.
I will explore the critical role of HPC in supporting Defence AI/ML research. I will provide an update on the development of our new HPC capability for Defence Research and National Security to be deployed at DST. A key advantage of HPC technology is that it will dramatically lift researcher productivity and create new research opportunities. Research that would take years on a desktop environment can be done in hours or minutes on a modern HPC system. A modern HPC system supports traditional modelling and simulation tasks, machine learning and AI, and data analytics at massive scale and I will provide defence relevant examples of all these use cases. In conclusion, a modern HPC system will help to ensure that we maintain our competitive advantage in what is a highly competitive environment.
Biography
Prof. John Taylor is currently Research Group Leader in CSIRO Data61, and Chief Computational Scientist, at the Defence Science and Technology Group and Honorary Professor, College of Engineering & Computer Science at ANU. He leads complex, multi-site, large scale interdisciplinary teams of research scientists, computational scientists, computer scientists and software engineers that are delivering high quality strategic science. He has held leadership positions managing large and diverse programs of research and teaching at prominent universities and research laboratories in both Australia and the United States of America. In these positions he has taken the lead in developing the vision, the culture of excellence, setting the strategic directions, building high performing teams and delivering on the strategic goals. Prof. Taylor developed one of the first HPC/GPU clusters at CSIRO in 2009 achieving a ranking of #146 on the TOP500 list in 2010. CSIRO has operated HPC/GPU clusters continuously since 2009. Lately, I have lead the development of a major new HPC capability for Defence and National Security research.
