Professor Attila Brungs gave the opening remarks at Supercomputing Asia Conference 2024 in Sydney.This was the first time the conference had been held outside Singapore, making the immense potential for regional collaboration to address some of the important scientific challenges of our time
I start by acknowledging the Gadigal People of the Eora Nation, the traditional custodians of the land on which we meet.
I pay my respect to Elders, past and present, and I extend that respect to Aboriginal and Torres Strait Islander people who are here today.
Well, what an honour to speak to you this morning – at the very first Supercomputing Asia Conference to be held outside Singapore.
Australia has some of the world’s leaders in supercomputing fields including quantum and artificial intelligence – and we are fortunate to host world-class supercomputer data and science hubs like NCI Australia, in Canberra.
Hosting Supercomputing Asia here in Sydney is a great symbol of how we can collaborate, right across our region, to realise the enormous potential of exascale computing and address critically important scientific challenges.
Challenges like climate change, environmental sustainability, medicine and healthcare – these are shared, human problems – and supercomputing can help us solve them, together.
I would like the thank all the organising institutes for creating such a golden opportunity for conversation and cooperation:
These institutes are enabling our scientific and research community to push the boundaries of scientific exploration – to improve the lives of people right around the world.
And that’s one of the reasons why UNSW Sydney is proud to be a ‘Major Collaborator’ of NCI Australia, and why I am very pleased to be opening this conference.
Our vision at UNSW is to improve lives globally – having a positive impact on the world around us drives everything we do – our education and our research.
At UNSW, we see pushing the boundaries of scientific exploration as a definitively human pursuit.
Society today is mechanised and digitised. Successive revolutions in agriculture, industry and communications have created an ecology where human ingenuity and autonomy are augmented by artificial intelligence.
We have self-piloting cars, trains, boats and drones, we use computers that automatically trade our stocks, and we use Google Assistant or Siri to manage our diaries, make phone calls, and check the weather.
Each day, with each new breakthrough in science and technology, it is becoming clear that we are racing toward a future with immense potential to drive productivity and improve standards of living across our community.
Supercomputing – HPC, AI, quantum – these are all part of a never-ending journey of discovery.
And, as has always been the case with scientific exploration, finally pushing those boundaries out always begets new boundaries to push – new frontiers to explore.
Making new ‘possibles’ drives us to discover and explode the next ‘impossibles’.
Such is the case with High-Performance Computing and Data.
That’s what makes ‘Exascale’ – the next frontier of supercomputing – truly mind-blowing.
Exascale promises computing power a thousand times greater within a few years – making the realm of what will become possible almost impossible to imagine.
This is the shoreline of the vast ocean at which we stand.
Here we are, deep in the Fourth Industrial Revolution, where cyber-physical systems – intelligent computers with massive processing power – are already allowing massive datasets to be analysed to solve some extraordinary problems.
And we’re not speaking of only abstract problems that seem removed from human experience – you and I know that supercomputing power is delivering more precise modelling and forecasting in very applied fields such as climate science, medicine and health, and transport and traffic systems.
(Though I do wish you would hurry up – it took an hour longer to get here this morning because of a traffic snarl!)
With such immense computational power, the potential productivity gains are immense – in both the private and public sectors – across industries, and across disciplines.
Not only is it possible to undertake functions and transactions more quickly and efficiently than ever before, but connectivity continues to improve our capacity to work together across regional geography.
Add to this the – importantly – increasing environmental sustainability of supercomputing – some of the infrastructure in Australia is already powered with 100% renewable energy – and there are exceptional productivity dividends to recover.
Yet, in order to realise this potential, it is crucial that our education and research sector is adequately resourced, and is appropriately flexible – to ensure that people have the requisite skills to thrive in a rapidly changing world.
But, as Andrew Stevens, Board Chair of Industry Innovation and Science Australia, said recently, “The way scientists are using high-performance computing facilities to respond to global challenges is itself rapidly evolving.”
If we are to optimise its real scientific and productivity potential, it’s important to acknowledge that keeping state-of-the-art supercomputing capabilities is a global race.
It’s a race for sovereign problem-solving capacity.
It’s a race for frontier scientific discovery.
It’s a race for global competitiveness.
And it’s a race to secure a strong future for science and prosperity in our region.
And so, our immediate challenge is to make sure that we keep pace with the international rate of change.
This challenge is at the core of this very conference: ‘Exascale Readiness in AI, HPC and Quantum’.
One of the most important considerations in this race – and the way to realise potential productivity gains – is to invest well.
Here comes an understatement: Exascale supercomputing infrastructure is expensive.
It’s estimated that to stand-up and operate an exascale machine over five to seven years would cost more than one billion US dollars.
The good news is that a recent Hyperion study showed the economic return on investment is 44 times – but we’d need the one billion dollars first, right?
Large global economies may be capable of making that investment – but that kind of investment is virtually impossible for small and medium-sized economies – like many of ours – to consider alone.
So, what can we do to make sure we’re not left behind?
A number of smaller nations in the EU, and some private partners, have joined forces to create ‘EuroHPC’.
Together, they’re developing “a world-class supercomputing ecosystem in Europe”.
It’s an idea well worth investigating in the Asia-Pacific region.
As I’m sure you know, one of the purposes of this conference is to pursue the conversation about “a shared high-performance computing and Artificial Intelligence ecosystem, for both the public and private sectors, in Asia.”
Discussions about shared regional exascale capability will be ongoing, and I encourage you to engage well over the next few days and beyond, in a spirit of partnership and scientific endeavour.
Of course, advanced infrastructure demands advanced workforces.
We need to make significant investment in making sure our people have the right skills for the task at hand.
While some commentators voice concern that the world’s workforces will be deprecated by advances in supercomputing and this ‘Fourth Industrial Revolution’, the opposite may actually be true if we consider wise investment in human capital.
Indeed, inherent within this ‘computers will take over the world’ narrative is an implicit assumption that education holds less value that it once did. If AI has supplanted all jobs in the future, then what is the point of acquiring skills throughout one’s life?
This narrative is problematic for a number of reasons.
First, history is littered with futurists mistakenly claiming that technology will eliminate employment opportunity. In the Depression-beset 1930s, individuals warned that robots would decimate factory jobs. John Maynard Keynes, no less, forecast that technological progress might allow a 15-hour workweek, and abundant leisure, by 2030 (if only!).
Second, education is not a teleological progression of skills advancement. It is a process not a product – and the purpose of education is not to equip people with specific skill sets in targeted subject matters – but rather to ensure that people are productive citizens who are flexible and adaptable throughout their working lives.
As John Adams astutely noted, “there are two types of education… One should teach us how to make a living, and the other how to live.” A good education system manages to do both – throughout a whole lifetime.
Too frequently we take a Fordist view of education – in that we silo education into sectors and then expect each sector to function in isolation – churning out graduates who are identical cogs to serve a specific part of the economy. To genuinely realise the potential of computing across the workforce, this must be reversed. We must view education and skill acquisition as a multidisciplinary, lifelong pursuit that enables workforce dynamism.
Such significant change in workforce skills requirements is one of the reasons UNSW focuses so hard on building the skills for our undergrads not just for their first job but their career, and critically, giving them the ability to keep on learning – and why we have now expanded our educational offering into lifelong learning.
Knowledge and skills are no longer a one-off educational experience that comes through a single bachelor’s degree acquired to achieve an endpoint at graduation.
Education and skills development continues for a lifetime – such is the pace of change, and UNSW is committed to meeting that need in business and individual settings.
To bring simplicity to an extraordinarily complex field: if we have advanced supercomputing capabilities and an appropriately skilled workforce, we can overcome traditional limitations of scientific exploration – relating to scope, scale and time.
As NCI sets out, supercomputers enable us to conduct experiments that are too small, too large, too short, too long, too expensive or too dangerous.
I’d like to share a few examples of how research colleagues at UNSW are using supercomputing to push the boundaries.
For example, Professor Katrin Meissner at the ARC Centre of Excellence for Climate Extremes, used a high-resolution atmosphere model to find a special kind of clouds that can develop under high concentrations of carbon dioxide.
These clouds have the potential to warm the high latitudes by several degrees and are still unaccounted for in climate projections.
Professor Jason Sharples, from the Bushfire Research Group based at UNSW Canberra, uses national supercomputing capability to investigate dynamic bushfire behaviour and the factors that lead to extreme bushfire development.
Jason’s research has led to the discovery of new forms of bushfire propagation and significant insights into how fire-atmosphere interactions affect the development of the most destructive manifestations of bushfire.
Economics Professor Loretti Isabella Dobrescu and her team significantly advanced understanding of what drives people’s complex decision making.
They were able to unpack, for example, how and why female leaders boost the performance of female students in higher education.
They used the first dynamic structural model dedicated to the function of education production.
Martina Lessio, a Senior Lecturer in Chemistry, uses national supercomputing resources to remove toxic arsenic from water with a novel class of materials, by unravelling their chemical mechanisms.
These are only a few examples, but I think they demonstrate three critical understandings about supercomputing.
One. Supercomputing is critical for scientific discovery and problem solving.
Two. Supercomputing has application across a breadth of disciplines and geographies.
Three. Supercomputing is highly technical but deeply human.
It’s this humanity that is the crux of this conference.
It is indisputable that humans’ ability to continue to make scientific discoveries – and to make the breakthroughs that change lives – will be increasingly dependent on our capacity to compute and analyse huge amounts of data.
As I said, we are at the shoreline of a vast ocean.
Are we ready to embrace the potential of exascale in AI, HPC and Quantum?
Are we ready to realise the productivity potential of supercomputing and this Fourth Industrial Revolution?
Are we ready to compete in the global race towards state-of-the-art supercomputing?
Are we ready to collaborate even more effectively across our region, for collective gains that benefit communities throughout the world?
These questions are as simple as they are complex.
I know you will explore these questions and so many more throughout this conference.
And I hope you thoroughly enjoy the opportunities to share your ideas, to learn, to inquire, to meet new colleagues and make new friends in this extraordinarily talented community.
Finally, I do believe that “innovation is a contact sport” – so please, bounce your ideas off each other, share your thoughts and experiences – this is where big ideas germinate and grow.
It is my great pleasure to open Supercomputing Asia Conference 2024.
I wish you a wonderful stay in Sydney.
