A brain scanning experiment at University College London (UCL) could help to create computers that make decisions in the same way as humans, heralding huge potential benefits for businesses.
Such computers would be able to identify trends and make accurate decisions about complex information much quicker than people, potentially replacing human participation in data analysis procedures.
The project, conducted at the Gatsby Computational Neuroscience Unit of UCL, used a functional Magnetic Resonance Imager (fMRI) to scan volunteer’s brains as they played a computerised gambling programme.
Dr Nathaniel Daw, one of the project’s leaders, presented subjects with a series of slot machines that paid out different amounts of money, leaving them to find the best one.
‘If we can understand how people solve problems using past experience we can design better decision-making machine algorithms that could be used in something like an autonomous robot, or in perfecting systems such as those used by Amazon.com to price books with,’ said Daw.
Understanding the brain as a machine from a computational point of view could completely change the capabilities of IT systems, says Michele Bezzi, a scientist at the Accenture Technology Lab. ‘For business, this could be very important, because it has a lot of decision-making tasks that are very hard for machines to do at the moment,’ he said.
‘While machines find it very easy to handle huge amounts of information, it is very difficult for them to make any sense of what any of it means.’
If human intelligence can be mimicked and applied to computers, machines would be able to think about the data they are receiving, identify patterns and make decisions for a business.
Simple approximations are already possible, says Bezzi, citing the example of Accenture’s own work on an intelligent video surveillance system.
‘The problem with surveillance is that there are often lots of cameras, but usually only one person to watch them, and that person could miss something or look the wrong way at a certain moment,’ he said.
‘An intelligent system could detect abnormalities and suggest to a guard that something wrong is happening.’
BT futurologist Ian Pearson says this is only the tip of the iceberg, and that computational neuroscience could lead us to machines with self-aware artificial intelligence as early as 2015.
‘Understanding how the brain works is going to give us a lot of clues as to how you might tackle making a machine conscious,’ he said.
‘If self-aware machines become commonplace, why would an organisation employ hundreds of employees to do eight hours’ work a day, when they could have a computer working flat out 24/7 and doing the job better than a human for a fraction of the cost?’
Further Reading:
Neuroscience
...in 30 seconds
- The field of computational neuroscience blends computer science with brain biology and applied mathematics to try to find out more about the workings of the brain.
- Scientists look on the human brain as a highly complex computer – one that is still only partially understood.
- Computational neuroscientists conduct experiments looking at how different areas of the brain respond to a variety of outside stimuli, such as pleasure or pain sensations, or what parts of the brain are used to tackle complex problems.
- The area is not just aimed at advancing IT research, but has medical implications for people with brain-damage or who suffer from mental impairment.
- By studying the responses of undamaged human brains, scientists can then compare the information with the activities of a damaged brain to determine what has gone wrong.
reader comments