When the University of Adelaide announced plans for the new, multi-million dollar Australian Institute for Machine Learning in December, it was seen as a great opportunity for the University to lead further development of this exciting, emerging technology.
Machine learning gives computers the ability to learn without having to be specifically programmed, allowing for everything from driverless cars to speech recognition and more effective web searching. It includes the still rather futuristic sounding concept of artificial intelligence, but is broader in scope – and is a reality right now.
According to alumnus, Professor Anton van den Hengel, the announcement of the new, multi-million dollar Australian Institute for Machine Learning (AIML) was a “perfect result” – for the University and for this exciting but still emerging area of technology.
“We’ve been widely recognised for our expertise for quite a while and it is a natural extension of that to set up a dedicated and very focused research centre,” Professor van den Hengel said. “We know there are a number of key areas, such as visual questioning answering, where we are the best in the world.”
“Machine learning has already changed how we shop, Facebook has changed how we communicate and Google has changed how we search for information – and they are just the easy things. It’s starting to have a big impact in agriculture and in mining and is about to change transportation,” Professor van den Hengel said.
“Machine learning will have a transformational impact on just about anything you can think of. The core technology gets applied all over the place and that’s the opportunity. It’s exciting. People are calling it the fourth industrial revolution.”
At last year’s announcement, the University’s then Interim Vice-Chancellor, Professor Mike Brooks said the AIML would help provide solutions to improve productivity, efficiency and service delivery for South Australians and attract globally dominant companies to the state.
“Through the institute, South Australia has the opportunity to lead the nation in the development and implementation of a strategy to create a vibrant ecosystem of high-tech businesses and highly productive workers,” he said.
The South Australian Government shares that sentiment. It has made a $7.1 million investment in the AIML, including $1.5 million prioritised for defence capability research.
The new institute will be one of the first tenants in the innovation precinct being established on the old Royal Adelaide Hospital site, occupying what was previously the Women’s Health Centre building on the corner of North Terrace and Frome Road.
Machine learning overlaps or complements fields such as computational statistics, mathematical optimisation, data mining and data analysis. The University has had strengths in these areas for many years bringing them together in 2007 with the creation of the Australian Centre for Visual Technologies.
In a decade, the centre grew from six people to more than 60, developing some important commercial relationships, not the least of which were in the defence sector.
“That was largely a computer vision group. Our transition to the AIML has seen us expand into other types of data and take on bigger projects and bigger challenges,” Professor van den Hengel said.
“Our skill sets are quite focused; we do machine learning technology. To use a motor vehicle analogy, we are experts at spanners and hydraulic jacks and we can apply them to all sorts of areas.
“We are working with agricultural companies to apply them to grow better crops and avoid common problems, we are working with defence companies on questions surrounding meaningful cooperation with robots, and we are working with medical companies making devices that will allow people to operate a powerful medical device on their own without medical training.
“Our main driver is that we want to do fantastic research. Some of that is theoretical research, but we also want to do high-impact research where much of that impact will be in other fields.”
And that means a number of collaborations are happening across the University, or with other research institutions such as the SA Health and Medical Research Institute (SAHMRI).
About half of the AIML team is research students and its Director hopes to expand those numbers. He also sees the potential to develop a specific machine learning curriculum at undergraduate level to encourage undergraduates to pursue postgraduate work in the field.
It is very much the way of the future.
Professor Anton van den Hengel is the Director of the Australian Institute for Machine Learning (AIML), former Director of the Australian Centre for Visual Technologies, a Chief Investigator of the ARC Centre Excellence on Robotic Vision and a Professor of Computer Science at the University of Adelaide.
Professor van den Hengel leads a group of over 60 researchers working in computer vision and machine learning, and has had over 250 publications, nine commercialised patents and more than $50 million in research funding. He has received a number of awards, including the Pearcey Award for Innovation and the CVPR Best Paper Award.
Story by Nick Carne