They're robots, but not as we know them

Robots with genes and robots that can change their shape were among research presented at the 15th ICONIP (International Conference on Neuro-Information Processing), hosted by AUT University's Knowledge Engineering and Discovery Research Institute (KEDRI), in Auckland last week.

Applied in real life, neural networking can be used in many areas, for example to solve problems in medicine and cyber-security, and to build more intelligent robots, says Professor Nik Kasabov, director of KEDRI, which researches artificial intelligence and neural networks.

On a very basic level, the neural networking field is about understanding the brain as an information processing machine, and using the knowledge of the human brain to build more intelligent computer systems, says Kasabov, who has published more than 400 papers, books and patents in knowledge engineering, intelligent systems, neural and fuzzy systems, bioinformatics and neuro-informatics, speech and image processing.

Thirty speakers from all over the world descended on New Zealand last week to talk about different topics in the multi-disciplinary area of neuro-information processing, says Kasabov.

One of the highlights of the conference was a presentation by Dr Kenji Doya, of the Okinawa Institute of Science and Technology in Japan, which showed neuro-genetic robots, says Kasabov.

"Robots are now not only based on fixed rules about how to behave, they now have genes, similar to human genes, which affect their behaviour, development and learning," he says.

Discoveries on how genes relate to learning have enabled the development of robots with genes, he says.

Another interesting presentation was the German Honda Research Institute's talk on the co-evolution of the brain and body in robots, says Kasabov.

"Robots no longer have fixed shapes. They can evolve, in a similar way as [humans] evolve," he says.

Genetic information, in combination with some interaction with the environment, enables robots to change their bodies. The Honda institute is working on getting machines to imitate animals -- a first step down a very interesting path, says Kasabov.

Neuro-information processing could also be used to cure brain deceases such as Alzheimer's, epilepsy and Parkinson's, says Kasabov. One presentation at the conference showed a way of curing epilepsy through a computational model of information from the brain.

Another presentation concerned curing Parkinson's through understanding information frequencies of the brain and emitting frequency signals to control the brain, making the Parkinson tremor stop immediately.

Another area where neuro-information processing is useful is in cyber-security. KEDRI, together with a research institute in Japan, is working on a project which involves finding out how to use neural networking to detect novel intrusions, spam and attacks, he says. While it doesn't have anything to do with the brain, it uses the mechanism of how the brain captures new information, he says.

Join the newsletter!

Error: Please check your email address.
Show Comments

Market Place

[]