Young researchers get Marsdens

5 November 2015

Eibe Frank

Computer scientist Associate Professor Eibe Frank has been awarded $410,000 to study alternative approaches to deep learning - the process of extracting complex features from data.

Sports in sites of war and disaster, time-of-flight photography, tree-based learning algorithms, and ancient aquaculture are University of Waikato projects that have received 2015 Marsden funding, announced today.

Total funding from the Royal Society of New Zealand for the four projects is $1.3 million over three years.

Dr Holly Thorpe received a Marsden fast-start grant of $300,000 to examine young people’s engagement with informal sports to improve their own and others’ wellbeing in war-torn and post-disaster geographies. Her four case studies will be Afghan youths' participation in skateboarding; a grassroots parkour group in Gaza; post-earthquake Christchurch; and post-Katrina New Orleans. Dr Thorpe’s colleague at Waikato, Associate Professor Belinda Wheaton, will also be working on the project.

Electronic engineer Dr Lee Streeter also gets a fast-start grant of $300,000 to re-engineer a new camera that can simultaneously image distances and velocities, to enable real-time accurate measurement of dynamic scenes. He’ll be working with Dr Gordon Wetzstein from Stanford University in the United States on the project.

A third fast-start grant was awarded to Tauranga-based marine scientist Dr Phil Ross. He’ll be combining archaeology and molecular ecology with Matauranga Māori (Māori knowledge) to examine Māori oral histories alongside archaeological records and toheroa population genetics, to gain a better understanding of the extent to which early Māori manipulated their marine environment. This research challenges the idea that species distributions are largely a consequence of natural processes and recent human activity. Dr Ross’ associate investigators will be Dr Huhana Smith from Manaaki Taha Moana and Dr Bruce McFadgen from Victoria.

Associate Professor Eibe Frank, a computer scientist, has been awarded $410,000 to investigate alternative approaches to deep learning, which refers to the process of extracting complex features from data. Deep learning using artificial neural networks has achieved record-breaking accuracy on challenging classification problems from computer vision, speech recognition, and natural language processing, but deep neural networks are very challenging to train successfully. Dr Frank and associate investigator Professor Bernhard Pfahringer are proposing to develop more user-friendly alternatives by extending their research on tree-based machine learning algorithms. These techniques will be made available through the open-source software WEKA, which was developed at the University of Waikato.

University of Waikato’s Deputy Vice-Chancellor Research, Professor Bruce Clarkson, says it’s very pleasing to see that several newer appointees have been successful in gaining fast-start Marsden grants.  He says this shows they are on the right trajectory for a successful research career and it’s pleasing to see this early reward for their strong performance.

The Marsden Fund supports excellence in leading-edge research in New Zealand. Projects are selected annually in a rigorous process by ten panels who are guided by the opinions of world-leading, international researchers. Funding is usually spread over three years for each grant.

There are two types of grants: Fast-Start grants worth $300K over three years for early career researchers and Standard grants that can be worth up to $850K for three years.

The Marsden Fund is contestable, is for investigator-driven research projects, and is not subject to government priorities. It is administered by the Royal Society of New Zealand and funded by the New Zealand Government.