Will computers ever be able to think like people? Professor Nikola Kasabov, a world expert in artificial intelligence, looks at what is likely to happen in the next 10 years.
"Artificial intelligence (AI) started with the idea of building intelligent machines that could think and communicate like humans. That still hasn't been achieved but along the way the concept of AI has been broadened. In the last 10 years AI developments have included data mining, knowledge engineering and discovery, bioinformatics, neuroinformatics, multimodal information processing (for example speech and images), which has all lead to the term 'computational intelligence' (CI).
At AUT's Knowledge Engineering Discovery Research Institute (KEDRI) the aim is to take AI and CI to the next level by making systems more humanlike', by mimicking the brain at its different levels of operation - genetic, neuronal and cognitive. Artificial neural networks (ANN) are used everywhere - you just don't see them. They're in software systems that predict weather and movements in stock markets, recognise your voice on the phone, control robots and trains, or discover genes and new drugs.
In the next 10 years AI and CI systems will become even more humanlike with new discoveries made about how the brain works and new types of ANN developed. It will be possible, for example, as part of these networks to model the interaction of euromodulators, such as dopamine and serotonin, in relation to emotions and cognition.
Emotions are important because without motions we can't make decisions. This research will have an impact on modelling brain diseases, such as Alzheimer's and clinical depression, along with facilitating the creation of artificial cognitive and decision support systems that will have more human-like cognitive and communication abilities. We also believe that the networks will become more 'conscious' and 'aware' of themselves and the environment. We are currently carrying out work in this area with researchers from Europe and Japan.
KEDRI, which is associated with the School of Computing and Mathematical Sciences in the Faculty of Design and Creative Technologies, is also applying quantum principles to create new paradigms of probabilistic information processing methods, which will be much faster and more accurate than current information processing methods. These advances in information processing combined with the increased 'intelligence' of computers will enable researchers to develop a host of new applications. Applications might include personalised medicine, which takes an individual's unique profile into account when diagnosing treatment, or autonomous cognitive systems which could produce "thinking" robots that can interact with their environment and potentially extend independent living for elderly by providing support in the home."
Click here to find out more.