Thinking like a Human – Cognitive Computing

By Kal Kaur

Introduction
Systems of Neuromorphic Adaptive Plastic Scalable Electronics (SyNAPSE)
References

Introduction

One of the most fascinating research areas in automated systems is the possibility of making a computer think like a human brain. The challenge here lies in the ability to train a computer in a technological and physiological manner. But how would it ever be possible to incorporate the physiological changes that manifest in a human brain and mirror this in a computer system?

Imagine how the human body may display a certain reaction in response to an emotional stimulus, what you will normally see is a change in behaviour due to a change in thought process (e.g., the ‘fight-or-flight’ response – a primitive, automatic inborn response that prepares the body to fight or flee from a harmful situation). With this in mind, researchers will need to test the limits to computer processing and the human brain.

Systems of Neuromorphic Adaptive Plastic Scalable Electronics (SyNAPSE)

One particular milestone may be achieved by IBM corporations in the hope that ‘two minds meet as one’. This project named the Systems of Neuromorphic Adaptive Plastic Scalable Electronics (SyNAPSE) is being worked on by researchers at IBM with additional research teams from Columbia University, Cornell University, University of California, and the University of Wisconsin-Madison.

It is no question that the human brain is the most sophisticated computer system on this planet. Even if this six-year funded project by DARPA revealed a model close enough to what humans have evolved with, there is still the question of exactly how this chip would learn behaviour and age?

IBM have recently revealed how they have engineered two prototype chips that can process incoming data the way a human brain interprets and responds to external stimuli or physiological change. The chips are to power personals computers. According to the research team at IBM, the chips can demonstrate the need for computers to perform multiple tasks in a simultaneous fashion with varying input stimuli.

Why is this chip system going to be so important? A chip that, if comparing to a human brain, can process extremely large amounts of data and problem solve then it certainly becomes gold dust for corporations that require a system that can perform large data-crunching tasks.

You have to remember that we are talking about real-time input for this chip system and this is going to be one of the biggest hurdles for the research team to establish exactly how a computer will process simultaneous data as it receives it in an intuitive way - a trial and error technique that will involve sophisticated algorithms and silicon circuitry to move one step closer in training a computer to learn through experience. The following video explains how computer brains may rival the functionality of the human brain.

Neuromorphic chips to build systems that standard computers cannot do, such as object recognition could benefit the design and engineering of humanoid robots. It is about building a cognitive computer that can build large objects and machines and carry out problem solving tasks.

With this prototype chip there is a basic idea of an algorithm that will work to mirror how the network of neurons typically behave in the human brain. The step here is to take definitive number of algorithms and designing a hardware system that can run the algorithms in a real-time nature – a difficult goal to achieve. It will be fundamental to design individual neuronal synapses that can be gathered to create a neuron infrastructure that can be built into communication structure for the chip system.

The project is one of a kind and will hopefully provide a better understanding of the capacity to modern-day technology and how close we can get to designing the perfect brain.

References

IBM – SyNAPSE


Date Added: Sep 8, 2012
Tell Us What You Think

Do you have a review, update or anything you would like to add to this article?

Leave your feedback
Submit