Event Title: The best of both worlds: Supercharging simulations with AI using SmartSim
Date: 7 Dec 2023 4-5 pm AEDT
Description: 
Join our AI Month session featuring international expert Alessandro Rigazzi, ML & AI Engineer at Hewlett Packard Enterprise on the cutting-edge fusion of numerical simulations and AI with SmartSim—an innovative open-source project by HPE. Explore how this computational paradigm shift is transforming scientific research across diverse fields. Don’t miss the chance to hear from real-world examples where SmartSim has elevated simulation capabilities. Secure your spot now and stay ahead in the era of AI-driven simulations, register now.
For more AI events visit here
About the presentation
Numerical simulations, known as the third pillar of the scientific method, play a pivotal role across various fields, from fluid dynamics and molecular modelling to weather forecasting and Earth sciences. Recent advancements have led to a novel computational paradigm that integrates numerical algorithms with data-driven models. This approach not only enhances accuracy but also enriches the insights gained by researchers. SmartSim, an open-source project maintained by HPE, stands at the forefront of this evolution. It enables scientists to incorporate AI capabilities directly into their simulations, thus overcoming the limitations of traditional methods and achieving previously unattainable computational feats. This talk will showcase examples where SmartSim has been instrumental in pushing the boundaries of scientific simulation.
Host: ARDC ML4AU, Intersect 3AI Platform, Pawsey (Alessandro Rigazzi, ML & AI Engineer at Hewlett Packard Enterprise)
Location: Online (Zoom webinar)
Registration Link: The best of both worlds: Supercharging simulations with AI using SmartSim - Pawsey Supercomputing Research Centre

Event Title: “AI in Action: Real-World Research Applications and Impact of Artificial Intelligence for Image Analysis”

Date: 5 Dec 2023 3-4 pm AEDT

Description: Research Data Scientists will showcase concrete examples of how artificial intelligence is driving tangible results in various research domains. Gain firsthand insights into the practical applications of AI and its transformative impact on research across diverse fields. Specifically, the webinar will focus on the application of image analysis tools and their application within medical and ecology research.

Host: ARDC ML4AU, Intersect 3AI Platform (presenter), Pawsey

Location: Online (Zoom webinar)
Registration Link: https://intersect.zoom.us/webinar/register/WN_nV8gHs5vR7S35nr9oFYatQ

Event Type: Presentation and discussion

Date and time: Tuesday, November 28, 2023 |3:00 PM 4:00 PM|

Recording: This event may be recorded, and the recording may be published.

Registration: https://www.eventbrite.com.au/e/machine-learning-eresearch-platform-mlerp-tickets-746778633997

Location: Online

Zoom: https://zoom.uts.edu.au/j/82972888324 | Password: 101010

About the Event

To efficiently manage GPU resources and reduce entry barriers, Monash University, UQ, QCIF, and ARDC have implemented a machine learning service known as the Machine Learning eResearch Platform (MLeRP). They have also developed user-friendly training materials for easy adoption. MLeRP bridges the gap, offering the interactivity of a notebook and the computational power of an HPC environment, allowing resource sharing among users when code is not actively running. This platform is set to launch on November 7th and will be introduced by the Monash University team during this presentation. The team will delve into what sets MLeRP apart from other HPC platforms and how to leverage its hardware effectively. The presentation will include an overview of MLeRP's features, a live demonstration, and a brief introduction.

For more information about the platform, please visit https://ardc.edu.au/project/national-machine-learning-service. The presenters, Dr. Slava Kitaeff and Mitchell Hargreaves, represent all partner institutions but are from Monash University.

Event Title: “Productivity Reimagined: Can Technology and AI revolutionise productivity?”

Date: 21 Nov 2023 3-4 pm AEDT

Description: Explore the exciting possibilities of how technologies and artificial intelligence can transform the way we work and help us achieve peak efficiency. Discover a few key productivity methods and tips as well as innovative AI tools that can help you supercharge your productivity in an ever-evolving digital landscape.

Host: ARDC ML4AU, Intersect 3AI Platform (presenter), NCI?, Pawsey

Location: Online (Zoom webinar)

Registration Link: https://intersect.zoom.us/webinar/register/WN_t0X-sb57R12LFtBwzJuszA

Type: Presentation and discussion

Time: 4pm - 5:30pm AEDT

Zoom: Meeting ID: 829 7288 8324  |  Password: 101010

Description:

Talk 1: “Data Science Gap in Research"
Dr Isabel Ceron | 15 minutes
Isabel works as Senior Policy Analyst at Academy of the Social Sciences in Australia, a role that enables her to articulate the needs of the social scientists. Isabel is dedicated to finding implementable solutions to the challenges faced by Australian social science research. With a diverse educational background including a BA in Social Communication, MSc in Urban and Regional Planning, MSc in Regional Development, and a PhD in Science and Technology Studies, Isabel brings over 20 years of experience from across the media, government, consultancy, and non-profit sectors.
URL: https://www.linkedin.com/in/isabelceron/

Talk 2 : “Accelerate research by bridging the gap in AI and Data Science skills needed in modern research and upskilling the new generation of researchers”
Dr Anastasios Papaioannou | 35 minutes
Anastasios holds a PhD in Computational Biophysics/Medical Physics from the University of Sydney with his research focus mainly being on Computational Physics, Data Science and Machine Learning in applied medicine, biology, and physics. With over 10 years of experience in research, Anastasios possesses an extensive research background as well as a deep technical understanding and expertise on research data, data science, machine learning, data and business analytics, and programming. He works collaboratively with relevant stakeholders to develop and implement solutions to increase research productivity and advance their research. As Intersect’s Data Science Manager, Anastasios manages and oversees the services provided by Intersect’s Advanced Analytics & AI Platform and the Data Science Team. URL: https://intersect.org.au/person/dr-anastasios-papaioannou/

Q&A | 10 minutes
Meeting | 20 minutes
Wrap up | 20 minutes

Type: Presentation and discussion

Registration: https://www.eventbrite.com.au/e/ausnz-hpc-admins-for-machine-learning-anaconda-alternatives-tickets-349967300617

Time: 11AM AWST, 1pm AEST, 3pm NZST

Description: Anaconda Alternatives: How to setup python based ML software stacks without 'conda

Join us for our June event, where Oliver Cairncross will share their journey from 'conda to pip for ML.

"It can be difficult (or impossible) to create ML environments using Anaconda if package dependencies don't match infrastructure. When base libraries and drivers can't be updated or installed ad hoc to match a multitude of package requirements (i.e., HPC) things get painful. I've started to build these packages using pip to match given infrastructure and created wheels (pre built binaries) that could potentially be used by end users. I’ve only used this for my own work–no testing with general users. Nevertheless, I’m happy to discuss what I’ve covered so far; which includes:

• Cuda and ROCm (AMD) drivers;

• NCCL (Nvidia) and RCCL (AMD) GPU communication libraries;

• Various MPI flavors;

• The Gloo communication library; and

• The Tensorflow and Horovod frameworks."

Find out more on Friday the 24th of June at:

• 11am -12 noon AWST

• 1pm - 2pm AEST

• 3pm - 4pm NZST

Zoom details will be emailed as a reminder closer to the event as well, but if you're looking for them ASAP:

Join from a PC, Mac, iPad, iPhone or Android device:

https://monash.zoom.us/j/84479348652?pwd=b2xEeDJpc1hCQmgrQ01EMFczYVJFdz09

Meeting ID: 844 7934 8652

Passcode: 544407

Who are the AusNZ HPC Admins for Machine Learning?

We're a group of HPC system administrators and eResearch support staff who want to learn how we can support the growing community of researchers using Machine Learning techniques. We're aiming to meet monthly, with mixes of workshops, presentations, rant sessions, and more!

If that sounds like you...

• Come along to our events to learn more;

• Come hang out with us in the #hpc-admins-4-ml channel in the ML4AU Slack space;

• Or propose a talk/workshop/event of your own that you'd like to run by contacting me, Kiowa Scott-Hurley (kscotthurley@gmail.com)

Type: Showcase events (Zoom)

Registration: https://intersect.org.au/education/joint-research-showcases/2nd-showcase/

Description:

AI is rapidly revolutionising drug design and discovery, fuelled by the promise of dramatically accelerating a very costly and time-consuming process. Four peak presentations will showcase state-of-the-art research activity in this fast-evolving field. The internationally renowned scientists reveal how large, high-quality datasets coupled with machine learning analysis can play a vital role in the process of new drug design and discovery. The presentations will be of broad interest to scientists working in the field of HPC and Big Data in drug design and biosciences more broadly, from advanced graduate research student level, to early career researchers and beyond.

Registration Link: https://www.eventbrite.com.au/e/ausnz-hpc-admins-for-machine-learning-the-great-divide-tickets-329444115217

Date and Time: Friday 20th of May

• 11am -12 noon AWST

• 1pm - 2pm AEST

• 3pm - 4pm NZST

The Great Divide: Balancing Support Across the Bleeding Edge and the Long Tail

Join us for our next event where Eric Burgueño walks us through their experiences with supporting researchers across The Great Divide.

One one half of the divide, researchers are working on the bleeding edge of machine learning and using highly complex methods, with admins scrambling to keep up with their ever-evolving needs.

The other half of the divide houses the long tail of newbies, who need an extra helping hand to get started.

How's a HPC admin supposed to keep up?

Event Type: Discussion and Community Building

Cost: Free

Location: Online - Zoom

Time zones:

30th March 2022

• 3-4pm AEDT

• 2-3pm AEST

• 12-1pm AWST

• 5-6pm NZDT (sorry NZ)

Who can attend: Everyone, but aimed at HPC admins and eResearch infrastructure support

Contact details: kiowa.scott-hurley@monash.edu

Registration: https://www.eventbrite.com.au/e/starting-a-hpc-sysadmin-and-machine-learning-community-for-ausnz-tickets-272447847847

Slack invite: Especially the channel #hpc-admins-4-ml

https://join.slack.com/t/ml4au/shared_invite/zt-141aeritd-aUtwRjy8s55DFdSpj0hA3g

Host Organisation: ML4AU

Details:
As the eResearch community finds itself supporting more and more researchers jumping into the world of Machine Learning, HPC sysadmins and eResearch support staff are learning right along with them.

We want to provide a space for these admins and support staff to be table to tap each-other on the shoulder and ask those painful questions like:

• What benchmarks do you use for ML?

• What do you do about conda environments on your parallel filesystems?

• Do I really have to allow Jupyter notebooks on a cluster?

• How many coffees can you drink before a multi-GPU job submission starts on your local cluster? ...is there a better way of tracking this?

We also want to provide a good space for the usual nonsense and banter that arises when you get a group of us in one spot.

This first meeting will be focused at saying hello to all your new friends, and working to decide what the community looks like - do we want to run online catch-ups every month? Every week? Every quarter? Do we want to do knowledge sharing showcases? Would we rather just be Slack friends?

Let’s find out together.

Event Type: Seminar

Location: Online

Time zone: AEST (GMT +11)

Who can attend: Everyone

Registration: At this link

Host Organisation: AWS

Details:
The transformer is one of the most popular state-of-the-art deep (SOTA) learning architectures that is mostly used for natural language processing (NLP) tasks. Ever since the advent of the transformer, it has replaced RNN and LSTM for various tasks. The transformer also created a major breakthrough in the field of NLP and also paved the way for new revolutionary architectures such as BERT.

Join this session

• To dive deep into Transformers, and understand how it uses the encoder-decoder architecture for a language translation task.

• We will learn how transformers overcome one of the major challenges which we had in recurrent models like RNN/LSTM in capturing the long-term dependency. Later we will learn about BERT and see how it differs from other embedding models and build a Disaster Tweet Analysis system.