Memristor Technology Offers The Promise of Learning Computers

Thursday, April 11, 2013

memristor nanowires

Nanowire technology is being used to create 'memristors' which may change the way computing works by making it more similar to how our brains operate. Researchers have discovered that exposing a random network of nanowires to stimuli like electricity, light and chemicals, generates chemical reaction at the junctions where the nanowires cross.
In the contemporary world of computation, bit flips are the base of the mathematics that the digital computer performs — the 1's and 0's or on and off signals that take place in the microprocessors.

There is another system though that works more the way our brain does and it is rooted in memristor technology. Memristors were originally envisioned in 1971 by circuit theorist Leon Chua.   Leon Chua has said that the memristor is the oldest known circuit element with its effects predating the resistor, capacitor and inductor.

When current flows in one direction through a memristor, the electrical resistance increases; and when current flows in the opposite direction, the resistance decreases. When the current is stopped, the memristor retains the last resistance that it had, and when the flow of charge starts again, the resistance of the circuit will be what it was when it was last active.  Memristors can therefore combine the functions of memory and logic, much like the synapses of biological brains.

The technology is even thought to be a component that could allow sentient computing.

Related articles
Now John Boland, Director of the Centre for Research on Adaptive Nanostructures and Nanodevices (CRANN), has been awarded a €2.5 million research grant by the European Research Council (ERC) to further develop the technology.

The award will see Boland and his team continue world-leading research into how nanowire networks can lead to a range of smart materials, sensors and digital memory applications. The research could result in computer networks that mimic the functions of the human brain and vastly improve on current computer capabilities such as facial recognition.

Nanowires are spaghetti like structures, made of materials such as copper or silicon. They are just a few atoms thick and can be readily engineered into tangled networks of nanowires. Researchers worldwide are investigating the possibility that nanowires hold the future of energy production (solar cells) and could deliver the next generation of computers.

Boland has discovered that exposing a random network of nanowires to stimuli like electricity, light and chemicals, generates chemical reaction at the junctions where the nanowires cross. By controlling the stimuli, it is possible to harness these reactions to manipulate the connectivity within the network.

What Boland and his team want to find out is if they let random connections happen in a tangled mass of nanowires, will emergent phenomenon result.  In this way, they are treating the memistor nanowires much like an infant's brain, where the connectivity of the synapses is not yet well established.

This could eventually allow computations that mimic the functions of the nerves in the human brain – particularly the development of associative memory functions which could lead to significant advances in areas such as facial recognition.

Boland said, “This funding from the European Research Council allows me to continue my work to deliver the next generation of computing, which differs from the traditional digital approach. The human brain is neurologically advanced and exploits connectivity that is controlled by electrical and chemical signals. My research will create nanowire networks that have the potential to mimic aspects of the neurological functions of the human brain, which may revolutionise the performance of current day computers. It could be truly ground-breaking.”

SOURCE  Trinity College Dublin

By 33rd SquareSubscribe to 33rd Square

Enhanced by Zemanta