Navigating the information avalanche
How do you make sense of unstructured data? If a medical centre captures data for over 5000 patients in a random manner, how do you identify recurring patterns? Or how do you gather meaningful insights about earthquakes from a number of unlabelled records including variables like magnitude, depth and location?
Everyone collects data these days—companies, schools, sports teams, researchers, you name it. But what does it all mean?
“We need ways of investigating, displaying, transforming, summarising and simply making sense of this avalanche of information. Data Science and Artificial Intelligence provide us with the tools we need,” says Professor Richard Arnold.
Professor Arnold and Professor Bing Xue lead the University’s new postgraduate programmes in data science and artificial intelligence (AI), respectively.
Having started his career as an astronomical researcher, his time at the Universities of Cambridge and Leiden (The Netherlands) offered Professor Arnold a strong foundation in scientific observation, mathematical modelling and data interpretation. “I later moved into statistics and though they may seem largely unrelated, astronomy and statistics have a lot in common. Statistics is fundamental to the process of interpreting astronomical observations, gathering insights, and drawing conclusions from data.”
“My foray into computer science was more of a happy accident,” says Professor Xue. “I was pursuing a degree in management science and took a paper in data mining. It was all about finding patterns from huge datasets and I enjoyed trying to make sense of the larger picture—so much that I decided to switch fields.”
Professor Arnold became a research associate in Statistics at the School of Medicine, Epidemiology and Public Health at the Imperial College in London as his first step towards a career in statistics—“That research was all about environmental risks to health, including the possible toxic effects of landfills.” After returning home to New Zealand he worked at Statistics New Zealand, where he saw the relevance of statistics to the public sector. “I was later able to combine medical statistics with public policy when I was involved in evaluating a national vaccination programme for Meningococcal B, and evaluating the health effects of the fund assisting New Zealanders to insulate their homes.”
After completing her earlier undergraduate and Master’s degrees in China, Professor Xue moved to New Zealand to pursue her PhD at Te Herenga Waka—Victoria University of Wellington. “My interactions with the staff and the University’s reputation for high-quality research were the key reasons I chose to come here. I had a lot of support from the staff and other students, so I didn’t find it very difficult to make the transition.”
Through their research, which involves close collaboration with researchers from other universities and industry partners, Professor Arnold and Professor Xue aim to develop novel approaches towards addressing real-world problems in areas such as biology, biomedical sciences, and aquaculture.
“With research, there’s always a chance that you end up exploring unexpected avenues—and that’s one of the reasons I enjoy it so much” says Professor Arnold. “I was once approached by a geophysicist who wanted to know, from a statistical point of view, what could be learned from individual earthquakes regarding the tectonic stress that causes them. The interesting aspect of this for me was the highly geometrical nature of earthquakes—they are movements on tilted planes sometimes very deep beneath the earth’s surface.”
This led Professor Arnold along a sequence of further projects, including analysing data from the 2011 Christchurch and 2016 Kaikōura earthquakes, and more recently his work in geometrical statistics has led him into crystallography in which he studies the highly symmetric arrangements of atoms inside solid metals. Nowadays he works on methods of finding structure in data, such as the content of trawls from fishing vessels, and analysing the lifetimes of complex repairable systems such as aircraft.
The universal nature of artificial intelligence piqued Professor Xue’s interest in this area at a young age. “AI is everywhere today and it continues to evolve into an integral part of our daily routines. And I find that very intriguing and inspiring—the ability of technology to become such a seamless part of our lives.”
Professor Xue is currently researching evolutionary machine learning, which uses evolutionary computation methods to address various machine learning problems. These include feature selection, high-dimensionality reduction, classification, regression, clustering, image analysis, evolving deep neural networks, transfer learning, and multi-objective optimisation. She has received two prestigious Marsden grants for advancing research in this constantly-evolving field—one was for a project on classification accuracy, while the second focused on new evolutionary computation-based approaches to automated design of deep convolutional neural networks (DCCNs).
Professor Arnold and Professor Xue are currently part of a NZ-wide collaboration that has recently received a $13 million grant from the Ministry of Business, Innovation and Employment towards developing zero-carbon aquaculture through data science.