23 July 2013

Australian Research Council (ARC) Laureate Fellow, Professor Peter Hall, has been receiving respect and praise for his work in developing theoretical tools to determine the consequences of analysing data in Australia for many years. A recent phone call from a colleague in Washington DC revealed the international scale at which his work has now been acknowledged.

The prestigious US-based National Academy of Sciences (NAS) recently announced the election of 84 new members and 21 foreign associates from 14 countries in recognition of their distinguished and continuing achievements in original research; Professor Hall, who is based at the University of Melbourne, was named as one of those foreign associates.

“The news came as quite a shock. I received a phone call from Washington DC. An American colleague was on the line and I remember being surprised and saying, "It's very nice to hear from you, what I can do to help?"

Then he told me why he had called and I almost dropped the phone. I had not been expecting this at all,” said Professor Hall.

ARC Chief Executive Officer, Professor Aidan Byrne, congratulated Professor Hall on his election to the NAS and said the ARC was proud to be supporting his research.

“To be entered as a Foreign Associate of the distinguished NAS is a notable achievement. Not many Australian scientists have received this accolade and I congratulate Professor Hall on such an outstanding achievement.”

Maintaining research at a high level and working as hard at his research today as he ever did in the past, Professor Hall feels sure that the decision to be accepted into the notable establishment must have been based on his contributions to statistical science over a reasonable period of time.

“I started off working in probability theory, but I had difficulty finding permanent work in that area. Although I subsequently won recognition for my research there, it took a fair while to percolate through. In the late 1970s and early 1980s I began to do research in theoretical statistics, not least because I managed to get a permanent job at the Australian National University in that area. Essentially, that is what I've been doing ever since,” said Professor Hall.

Historically, in the late 1970s and early 1980s high levels of interactive computing power started to become readily available to statisticians working in universities. This motivated the development of particularly sophisticated methods for analysing data. Those methods required relatively few assumptions about the population  from which the data came.

“I've been working in theoretical statistics, applied to computer-intensive statistical methods, ever since the early 1980s. I was in the field almost as soon as it began to grow. My experience with theory has given me an excellent foundation for describing just what it is that computers are doing to data.”

Professor Hall’s early statistical research has secured him a number of ARC grants over the years to pursue his enquiry. He is currently working on a project, New directions, new problems and new data types in statistical science, funded under the ARC Australian Laureate Fellowships scheme.

“The funding has helped a great deal to give me the time I need to think problems through, and work out just what has to be done. My line of work requires a great deal of thinking time to develop a good research programme.

“With my current funding, my group and I have developed a remarkably accurate way of analysing data in the form of functions, using an adaptive approach that identifies very high-dimensional aspects of those data. For example, in the case of data functions, such as spectra, which vary over a continuum, much of the information being sought is often in relatively high-dimensional components of the curves. We have developed an algorithm that locates those features in an optimal way.

“We are also addressing problems where data are not observed directly, but indirectly with a significant amount of error. Here it is possible to solve a large class of problems under the most rudimentary assumptions. Moreover, the solutions are remarkably accurate. This surprised us a great deal. We are still trying to understand why the method works much better than we had expected,” said Professor Hall.

Professor Hall acknowledges that collaborative work has been crucial in making progress. The fellowship has allowed him to attend international conferences to meet with specialists in the field and bring these collaborators to Australia to progress the work further.

“I have good theoretical skills, and indeed theory is my passion, but I need much more, particularly these days, to make progress. For example, skills in statistical computing and in my case those skills are not nearly as strong as I would like them to be. So working collaboratively is essential.

“I am very fortunate to have the opportunity to work with excellent people both in Australia and in other countries around the world. These collaborators are themselves working very hard to make the sort of advances that give us a great deal of pleasure.

“This area of research is a very vibrant field of science that never stands still. It is always reinventing itself to handle new scientific problems, new types of data, and new levels of computing power. I am really fortunate to have been able to live and work during such an exciting period, and to still be experiencing excitement in my work,” said Professor Hall.

Professor Hall hopes to attend the ceremony marking his entry into the Foreign Associate of the National Academy of Sciences in April 2014. The event will be held in Washington DC.

For more information please email Professor Peter Hall at the department of Mathematics and Statistics at The University of Melbourne.