How Computer Brains Learn Human Reasoning
Source: Adrian Bridgwater
IBM Watson for President? No not quite yet… but ‘inductive reasoning’ and learning exhibited by computer brains may one day outstrip human intelligence. Image Source: Watson2016
As a piece of terminology, cognitive computing describes systems that are capable of learning and reasoning with both purpose and intellect. Cognitive systems are capable of interacting with humans ‘naturally’ i.e. in a manner that we find as almost human in many ways.
So could computers become smarter than humans one day soon? Should we hand over presidential power to some computerized inductive reasoning engine that evaluates all of societies troubles and challenges and always selects the most algorithmically correct outcome?
To answer the question, first we need to understand how computer brains learn.
The ‘birth’ of a computer brain
Although we do still need software application developers to build the code that runs cognitive computing systems, they are not ‘programmed explicitly’ in the same way we build traditional software. Instead, we provide cognitive systems with a framework of intelligence that they can add to as they learn and reason from interactions with humans and from their ‘data experiences’ with other devices inside their prescribed environment.
Essentially then, cognitive computing systems are ones that have progressed from the deterministic systems of old and become probabilistic. This allows cognitive power to employ pattern recognition techniques and give these new computers the ability to reason and to learn.
| }
|