Computerworld

Creativity and computational intelligence

AUT's Nik Kasabov pushes computing boundaries through the Knowledge Engineering and Discovery Research Institute (KEDRI)
  • Rob O'Neill (Unknown Publication)
  • 06 April, 2010 22:00

Dossier

Name: Nikola Kasabov

Title: Professor of knowledge engineering; director of the KEDRI Institute (www.kedri.info)

Organisation: Auckland University of Technology

Location: Auckland

Favourite restaurant: Masala, Mission Bay, Auckland

Most recent read: Vila Pacifica, by Kapka Kassabova (a personal copy)

Favourite place to visit in New Zealand: Dunedin.

Worst job: Assistant lecturer at the Technical University in Sofia, Bulgaria.

First computer: IBM mainframe

What keeps you awake at night: New ideas.

Nik Kasabov is pushing the boundaries of IT through AUT’s Knowledge Engineering and Discovery Research Institute (KEDRI), while also developing the next generation of scientists equipped with new ways of thinking about computing.

You founded the Knowledge Engineering and Discovery Research Institute at Auckland University of Technology in 2002. What does it do?

KEDRI develops new methods and systems for information and data modelling, data mining, knowledge discovery and computational intelligence, and applies them to important real-world problems. The methods are inspired by principles from nature such as biological neural networks and brain functions, genetics, evolution and quantum physics.

What would you say KEDRI’s biggest achievements have been?

The biggest achievement so far has been its pioneering role in the development of new paradigms of computational intelligence. For example: adaptive systems in the form of evolving connectionist systems (resulting in a Springer monograph book and many journal publications); personalised modelling (resulting in a patent); gene and DNA data profiling (software systems); computational neuro-genetic modelling (another Springer monograph book, software systems, journal publications); quantum inspired optimisation algorithms (journal papers). All these methods and systems are widely accepted and used in many laboratories in New Zealand and in the world.

A second major achievement of KEDRI is the education of a large group of young scientists (15 PhD students and 20 Master students so far) equipped with new ideas and new ways of thinking in the areas of information science, computational intelligence, bioinformatics and neuro-informatics.

What are your priorities and goals for KEDRI now?

At present we are collaborating with scientists from Europe, Japan, China and the US and are applying together for international research grants, including FP7 European grants, the Human Frontiers Science Promotion Fund, along with applying for national Marsden and HRC funds to enable us to remain competitive with the rest of the world.

KEDRI has been funded by government, but also by private sector partners including Fonterra and Telecom? What commercial research does it do and how is this originated?

The commercial work we have done is mainly at the level of developing new information processing methods more efficient than the existing ones, and to demonstrate them with prototype systems and experimental results. Often improving prediction rates by just a few percentage points can result in millions of dollars profit and thousands of lives saved. We have done projects with Fonterra, Telecom, ViaLactia, PEBL, Middlemore hospital and others.

Where can you see neural networking research taking us?

Artificial neural networks (ANN) are computational models that mimic the brain in its main functions of learning, generalisation and synaptic plasticity. New ANN models are now becoming more biologically close, such as spiking neural networks (SNN). SNNs process information in the form of trains of binary spikes, similar to how the brain does it. They have a tremendous potential to be used for modelling brain data and brain diseases, such as Alzheimer’s disease, clinical depression and stroke. They can be also used for complex spatio-temporal pattern recognition tasks, such as recognising movements, predicting spatio-temporal ecological and environmental events, building autonomous learning and cognitive /robotic systems.

The next level of complexity for SNNs will be when gene information is included in the functional description of the SNN, leading to computational neuro-genetic models, as I suggested in 2003 and now becoming a hot topic.

Study of creativity is one area of KEDRI’s research and it is part of the school of design and creative technologies. What is going on in that research area?

Creativity is a fundamental concept, mainly studied from a philosophical point of view, but not much from an information or engineering viewpoint. We would like to look at this concept from different viewpoints and investigate research questions, such as: How is creativity manifested in the brain of subjects performing different creative tasks, for example writing a computer program, or writing an article, or drawing a picture, or creating music? How can information be represented/visualised to enhance creative thinking? How is creativity enhanced through communication?

One of KEDRI’s goals is to create intellectual property for the benefit of New Zealand. Can you give some examples of this?

One of our goals is indeed to create new IP for New Zealand and to protect it internationally and to enable New Zealand companies to take it further and commercialise it globally. I think commercialisation of academic IP is still a problem for New Zealand. An example is a new method for personalised modelling that we registered as a patent in 2008 and which is still waiting for commercialisation, despite the acute demand for such methods.

What are you working on personally?

I have just signed a contract to be the sole editor of the Springer Handbook of Bio- and Neuroinformatics — a giant project of 1500 pages, written by more than 150 authors. And I also intend to keep and even improve my accordion playing, swimming, yoga, and painting skills, and also to enjoy life.

— Nik Kasabov is Kapka Kassabova’s father