New innovation centre offers graduates place to turn big-data dreams into reality.

Mapping cities in 3D with drones, building computers like brains and turning data into a virtual world you can walk through: these are the mind-boggling concepts that Auckland's newest hub for tech research will make reality.

The collaborative Interact Centre of Technology Excellence, to be launched tonight by hosts Auckland University of Technology (AUT), aims to make New Zealand a leading developer of cutting-edge data tech, while helping retain the country's top IT and data science graduates.

By tackling big issues through big data, the centre's researchers will develop high-tech ways to improve driverless cars, city transport, social services, precision medicine, environmental monitoring and precision agriculture.

The centre - bringing together experts and researchers from five universities and other organisations - is expected to turn out a large number of new PhD graduates and encourage them to remain in New Zealand with post-doctoral fellowships and industry opportunities.


Professor John Raine, AUT's pro vice-chancellor of innovation and enterprise, said the centre would boost the country's use of data-heavy research and development - an area where we had lagged internationally, but which offered many industry benefits.

The development comes after the Government yesterday launched a $35 million contestable fund for universities to lure big-name researchers to the country.

The "Entrepreneurial Universities" programme involved the Government entering a 50/50 partnership with individual universities to attract and support named researchers and their teams for an initial period of three to five years.

Tertiary Education, Skills and Employment Minister Steven Joyce said while the country's universities already had a good reputation for excellence, there was a need to keep adding to our high-tech sector.

But innovation commentator and University of Auckland physicist Professor Shaun Hendy said that while the new investment might slow the decline for a short while, what would keep our rankings high in the longer term was a better funded and well functioning science system.

Making data into a walk-through world

Stefan Marks of the Sentience Lab is transforming data into an immersive 3D space.
Stefan Marks of the Sentience Lab is transforming data into an immersive 3D space.

When you think of data, you might just think of dry numbers - but what if you could see what they represented?

AUT's Sentience Lab, led by Dr Stefan Marks, is transforming data into a new, fully immersive and interactive 3D space that can be used for multi-sensory visualisation, education, modelling, animation, rapid prototyping and much more.

The lab combines modern technologies such as motion capture, virtual reality, head-mounted displays, electronics and wireless video transmission and a custom software framework that integrates all those components into one incredible solution.


Similar to Star Trek's Holodeck, Sentience Lab allows one or more users to simply "walk" through, experience, and interact with life-sized three-dimensional worlds, whether those are designed to look like reality or represent abstract information such as complex datasets.

Sentience Lab has a growing repository of applications, such as a real-time visualisation of artificial spiking neural networks, a visualisation of New Zealand earthquake data, anatomical models for medical education, a 3D sketching tool, a visualisation of global CO2 emissions and more.

How to map cities in 3D ... using drones

Could drones be used to create three-dimensional maps?

3D maps, generated using aerial imaging and computer graphics, are today a common way to visualise cities and landscapes. But their creation is still costly.

However, PhD student Hsiang-Jen Chien has found a way to create inexpensive updates of maps, with specially designed computer vision software. Video data captured at 30 frames a second by a camera on a drone can be analysed then built into the 3D model.

The results so far have been highly accurate, opening up the potential to drone-map streets, housing developments, changing cities and coastal erosion.

Solving the driverless car fear-factor

The death of a driver using the autopilot feature of Tesla's Model S may have cast doubt on the safety of driverless cars, but they're still widely expected to become commonplace in the future.

They still can't sense challenging surroundings with high accuracy. But by enhancing the car's stereo vision with cameras that compare different views to determine its position and identify anything that might cause a collision, the technology will be able to deal with more challenging situations.

AUT engineering professor Reinhard Klette is working with researchers and firms around the world and next year will trial a test vehicle with fully equipped sensors at Northland's Transport Testbed.

Can computers think the way we do?

Professor Nikola Kasabov of the Auckland University of Technology tests out his innovation, the Neu-Cube, on a student.
Professor Nikola Kasabov of the Auckland University of Technology tests out his innovation, the Neu-Cube, on a student.

In the exploding world of big data, we're in increasing need of smart, powerful technologies that can process it all. And when it comes to intelligence, there's no greater supercomputer than our own brains, which excel at continuous learning and recognising patterns.

Designing computers to think the way we do has been a big focus of AUT's Knowledge Education and Research Institute (KEDRI), which has developed a ground-breaking technology called NeuCube.

Specialising in machine learning, pattern recognition and predictive modelling, NeuCube is the world's first evolving spatio-temporal data machine that works with brain-like principles.

Its goal is to learn useful patterns and trajectories to perform tasks as varied as controlling robots and neuroprosthetics and predicting earthquakes.