Dr. Stephen Sanderson, The University of Queensland

Dr. Stephen Sanderson is a postdoctoral research fellow in Prof. Debra Bernhardt’s group at the Australian Institute for Bioengineering and Nanotechnology (AIBN) at The University of Queensland, where his studies focus on the theory behind computational methods for and statistical mechanics of non-equilibrium systems. Prior to that, he completed his Ph.D. at James Cook University under the supervision of Prof. Ron White and Dr. Bronson Philippa, where he investigated charge transport and exciton dynamics in organic semiconducting materials using kinetic Monte-Carlo and molecular dynamics techniques.

Title: Making the work-efficient parallel prefix sum do less work - Please find the video here.

Date/Time: May 31, 2022 11:00 AM

Abstract:  The prefix sum (cumulative sum) algorithm has been accelerated for parallel processing through various algorithms, including the work-efficient algorithm in which the calculation is performed in “up sweep” and a “down sweep” stages. This talk begins by discussing details of implementing and optimizing this algorithm on GPUs, as outlined in [1]. A common use of the prefix sum is to aid in selecting a random element from an array of probabilities. The cumulative sum of the array is generated, and then an element is chosen such that it is the last element with a value lower than or equal to a random number. In this talk, modifications to the work-efficient parallel prefix sum algorithm and the subsequent search algorithm are described which avoid unnecessary work in this specific “sum and search” use case.

[1] M. Harris, S. Sengupta, J. Owens; Chapter 39: Parallel Prefix Sum (Scan) with CUDA, GPU Gems 3 (2007), Available online: https://developer.nvidia.com/gpugems/gpugems3/part-vi-gpu-computing/chapter-39-parallel-prefix-sum-scan-cud

Dr. Jack Yang, The University of New South Wales

Dr. Jack Yang is a Lecturer in Material Studies with Artificial Intelligence at the School of Material Science and Engineering and Materials and Manufacturing Futures Institute at The University of New South Wales (UNSW). He is also a member of the Research Technology Service Team under the office of the Pro-Vice-Chancellor (Research Infrastructure) at UNSW, which provides support for computational research on High Performance Computers (HPCs) across the University. Jack obtained his BSc(Nanotech) in 2008 and Ph.D. in 2011 from UNSW. Before returning to UNSW in 2017, Jack was a Postdoctoral Research Fellow at Westfälische Wilhelms-Universität Münster in Germany, and the University of Southampton in the UK where he worked on developing new structure prediction and machine-learning methods for discovering new functional organic materials. Currently, Jack is leading his own group in AI-driven material studies at UNSW with major research interests in the electron and phonon dynamics in perovskites for electronic, photovoltaics, energy storage, and catalytic applications.

Title: Benchmarking VASP6 on Gadi - Please find the video here.

Date/Time: Jun 28, 2022 11:00 AM

Abstract:  Around 30% of the computational power on HPCs is currently devoted to running calculations using Density-Functional-Theory (DFT), an extremely powerful method that can help scientists to understand the thermodynamic and electronic properties of matter, being either molecules, nanostructures, crystalline or amorphous solids. Among all the DFT codes, Vienna Ab-initio Simulation Package (VASP) is one of the most widely used packages for performing DFT calculations on periodic systems. Recently, a major update to VASP (Version 6) has been released and is now accessible on Gadi. Whilst most of the new features introduced in VASP6 are based around accelerating post-DFT calculations (such as hybrid-DFT and GW approximations), standard DFT users will still be able to benefit from the new GPU-compilation of VASP6, which can provide a significant increase when running DFT calculations on larger systems. In this short seminar, Dr. Jack Yang from UNSW will present his benchmark results from running VASP6 on Gadi, focusing on features that one can see significant performance improvement in VASP6 compared to VASP5. We hope these new benchmark results will encourage existing VASP5 users to consider upgrading to VASP6, in order for them to make more efficient usage of the computational resources provided by Gadi from NCI.

Mr. Andrew Howard, National Computational Infrastructure

Andrew is the associate director, cloud services at NCI.

Title: Introduction to Nirin and ARE - Please find the video here.

Date/Time: Jul 26, 2022 11:00 AM

Abstract:

Nirin is NCI's Cloud Service. ARE is a web-based graphical interface for performing your computational research at NCI. It combines the familiarity of your regular desktop/laptop with the power of NCI’s world-class research HPC capabilities. ARE gives you access to NCI’s Gadi supercomputer, Nirin cloud and Data collections, all from an easy to use, graphical interface.

Dr. Sonika Tyagi, Monash University

Dr. Sonika Tyagi has a Ph.D. in Computational Biology and over 15 y of work experience in academia & industry. In 2018 she joined a teaching and research position at Monash University to establish her research program. She is currently a Machine Learning lead in the SuperbugAI flagship project. Her expertise is in developing new machine learning (ML)  tools and pipelines, and applying these methods to solve biological and clinical research questions. Tyagi has developed novel probabilistic models based on natural language processing (NLP) to predict non-coding RNA structures ab-Initio, and has implemented deep learning approaches to predict functional motifs. She applied these methods to successfully predict novel miRNAs and variants causing melanoma. Her current research focuses on integrative approaches to health informatics and genomics.

Tyagi was a chief investigator on an NHMRC funded project to computationally study human birth (2017). In 2020 she was also awarded an Early Mid Career Research (EMCR) fellowship by the Australian Academy of Science to develop AI models for the diagnosis of preterm birth. This was followed by an NHMRC success in 2021 for Targeted Delivery of Nucleic Acid Therapeutics for Preventing Preterm Birth (2022-25). She has received industry funding (2018-22), and University grants (2019-20) for developing AI resources to study genetic diseases. Sonika was one of the finalists for Women in AI (WAI) awards - Australia/Newzealand 2022 in the "AI in Health" category.

Title: Open Challenges in Applying Machine Learning across Digital Health and Computational Biology - Please find the video here.

Date/Time: Sep 27, 2022 11:00 AM

Abstract:  

Digital health refers to digital technologies that have direct or indirect impact on improving and monitoring public health and wellness. Computational biology involves using biological data to develop algorithms and models to understand biological systems and also their relationships with health and disease. Machine learning (ML) based methods already make an integral part of both digital health and computational biology research.

Our group works at the intersection of genome and healthcare data of multiple forms. We employ ML methods to automatically learn complex features from individual data types, and harmonise heterogeneous multimodal information. In this talk first, I will discuss computational biology methods we have developed to understand the grammar of biomolecules and their functions. Furthermore, I will cover our integrative informatics approaches to combine genomics with electronic health data to decipher the interrelationship of disease driving genomic and physiological factors. Throughout the talk I will highlight the challenges, limitations and opportunities of ML applications in digital health research and development.