Published on the 07/05/2021 | Written by Heather Wright
Partnership and education…
Academia and business will come together as one at the University of Waikato’s new Te Ipu o te Mahara AI Institute.
At least that’s the lofty goal of Professor Albert Bifet (pictured) who says he wants to make artificial intelligence accessible to everyone – and to have ‘huge’ impact for business and the wider New Zealand community, positioning the country as an international leader in AI.
In practical terms, the Institute is focused on helping translate New Zealand’s expertise in AI, real time analytics of big data and machine learning into commercial businesses and applications for the benefit of New Zealand industries.
“The main goal of the institute is to help businesses and organisations get the benefits of these new technologies,” Bifet, Te Ipu o te Mahara director, told iStart.
“Our main goal is to have impact.”
“Research into real-time analytics for big data offers huge opportunities to create new businesses and transform existing businesses in New Zealand. It offers a step-change in computer performance, the efficiency and effectiveness of processing the huge datasets behind deep learning, machine learning and AI,” he says.
“For me, the AI revolution is like the internet revolution. It is going to change how we do things for all business and all sectors. New Zealand needs to invest in that not to be left behind.”
Earlier this week the Australian Federal Government upped its AI play, announcing $53.8 million to establish a National Artificial Intelligence Centre within CSIRO’s Data61, with $34 million available for AI-based projects to address national challenges.
It has allocated $124 million for AI research and grants as part of a $1.2 billion digital package which will form part of next week’s budget.
Bifet says the AI Institute will have a strong focus on collaboration between academia and industry – something he believes is a crucial component to success, with research supporting entrepreneurship and the commercialisation of AI technology.
“Academia is very good at solving problems and industry is very good at implementing solutions for these problems.
“If we can collaborate together with research finding solutions and industry implementing them, then we [New Zealand] can be very, very strong on AI,” he says.
The institute is a continuation of the University of Waikato’s 25 years of work in AI and machine learning and it’s already got a strong track record in collaboration and commercialisation. Its WEKA open source machine learning library, a suite of Java-based sofware tools for ML and data mining, has been downloaded more than 10 million times. It’s the success of projects such as WEKA and the income through the sale of commercial licenses of the software, that enabled the University to install New Zealand’s most powerful supercomputer for AI applications – a NVIDIA DGX A100. And it’s that commercialisation that Bifet wants to see replicated through the Institute.
The Institute’s doors will be open for businesses who need assistance with AI projects and partnerships.
If a problem – or project – can’t be solved easily with existing offerings, the Institute will look at research and developing new solutions in partnership with the business.
“The nice thing of having this partnership is the university can develop the solutions and industry then can use and implement them,” Bifet says.
The University’s Waikato Link arm, which works exclusively on commercialisation of projects, will help with any commercialisation opportunities.
He points to one of the University’s most recent successes, a collaboration with parking management company Parkable, as an example of a successful commercial collaboration and one it would like to replicate with other businesses.
The Parkable project sees the University and Parkable prototyping a car monitoring system using ML in cameras, rather than sensors, to detect car parks, directing parkers to available car parks in real time and enabling car park owners to monitor occupancy. The system will process video data in real-time, monitor multiple parking bays at once and run on several cheap computers, removing the need for high bandwidth, expensive servers and cloud processing.
“Our main goal is to have impact,” Bifet says. “We are really keen on partnering with companies and organisations so they can really get a huge impact.”
He points to the University’s TAIAO (Time-evolving data science/AI for advanced open environmental science) project, a seven year initiative funded by MBIE to advance environmental data science by developing new ML methods.
“There we can have a huge impact in environmental science and then for climate change. This is something NZ could be very strong in internationally and we are very happy to push for it, and then work with all the organisations and companies that are related to environmental science.”
But the Institute’s focus won’t only be on collaboration and commercial opportunities, with a strong focus on increasing New Zealand’s AI skills base.
“We are working on launching a Masters in AI for next year,” Bifet says.
“We are very very keen on extending our degrees to support AI and also we are looking at courses specific for industry.”
Those courses will include one at an executive level to help executives understand what AI is, it’s benefits and what it can and can’t do.
Other courses will also be provided for line managers and for technicians.
“We really think the way to push for AI is to provide a good education,” he says.
The Institute is also looking at how it can provide courses to high schools.
“This is going to be very very useful for two things: iI’s going to allow them to understand what AI can do, but at the same time it will allow them to understand what the risks of AI are.
“Right now it is very important to look at what are the potential risks in terms of bias, discrimination, privacy and especially data sovereignty. That is something NZ can contribute because there is all this expertise on data sovereignty from the Maori culture and it’s also a very nice contribution from New Zealand to the world.”
That issue of data sovereignty from a Maori perspective is one area of research the University is already engaged in, along with how to deploy Maori AI and explainable AI.