Artificial Intelligence

As Deputy Head of the Newcastle Robotics Lab, Dr Alex Mendes gets to travel the world with the NUbots, the University's robot football team that competes each year at RoboCup, the Robot Soccer World Cup.

Dr Alex Mendes and RoboCup team In 2015, he will accompany a team of students and their robots to Hefei, China.

While the competition itself sounds like a lot of fun, there is a serious side to it as a testing bed for ground-breaking artificial intelligence and intelligent robotics research. The NUbots team's research vision is to develop and program robots that can support humans not only for routine, challenging or dangerous tasks but also to improve quality of life through personal assistance, companionship and coaching. This vision supports the Newcastle Robotics Lab's overall mission of contributing to the responsible development and application of robotics.

While Mendes' background and research focus on the areas of optimisation and data mining contribute significantly to realising the broader aims of the NUbots research, his work takes many other directions. It is now increasingly moving into the health arena through his association with the University's Priority Research Centre for Bioinformatics, Biomarker Discovery and Information-Based Medicine.

Recently, a conference paper co-authored by Mendes for the 12th Australasian Data Mining Conference (AusDM'14) presented a neuro-evolutionary algorithm for detecting breast cancer tumours from mammographic images in its early stages so that treatment has a better chance of success. The team plans to test the algorithm on other biomedical benchmark case studies, in particular Parkinson's disease. Other examples of Mendes' publications in that area include the analysis of brain electrical activity for the detection of childhood absence epilepsy; and the analysis of genome-wide microarray data for the detection of subtypes of breast cancer.

Mendes explains the applications of his work in the health arena: "In 2003, when I first worked in health research, most studies would gather genetic data for only a few hundred genes and a handful of patients – the technology for DNA sampling was very expensive and in its early stages. Now, a person's entire DNA can be sampled at the cost of a few hundred dollars, so studies generate an immense quantity of data – typically thousands of patients and their entire DNA, or 30,000-plus probes. To make sense of all that data, you need quite complex computational methods to separate noise from valuable information about the genetics of the disease under study and, from that data, create diagnostic kits and targeted cancer therapies."

But the real-world applications of Mendes' work don't stop there. He is currently collaborating on a project with the Hunter Valley Coal Chain Coordinator (HVCCC) to optimise coal transportation between Hunter mines and Newcastle harbour, using mathematical modelling and evolutionary algorithms. In addition, he worked in partnership with Ausgrid on computational methods to optimise the electricity distribution network around Newcastle, and reduce the time required to restore supply to customers after a blackout.

With a PhD in Electrical Engineering, with a specialty in Automation, from the State University of Campinas, Brazil, Dr Mendes joined the School of Electrical and Computer Science in 2003 and is now the Head of Discipline for Computer Science and Software Engineering. His work has been published in many renowned international journals and conferences and he has been a reviewer for the Australian Research Council since 2006.