Nov 19 2018

Supercomputer Graham is helping to uncover secrets of the brain

SPAUN is the world’s largest functional brain model powered by Graham, one of Canada’s most powerful supercomputer.

Terms like artificial intelligence and machine learning bring to mind machines that can think, yet scientists are still puzzled as to how the human brain makes judgements or triggers the simple act of blinking.

One neuroscientist from the University of Waterloo has made significant progress in deciphering how those basic brain functions might occur. Dr. Chris Eliasmith is the director of the Centre for Theoretical Neuroscience where the globally renowned SPAUN brain model was developed. It is the world’s largest functional brain model that simulates 6.5 million neurons and is used to understand how the human brain works and how neurons communicate with one another. The large-scale brain model’s capabilities are quite outstanding with a digital eye for visual input, a simulated arm used to draw its responses and the ability to pass basic elements of an IQ test.

“This increase in speed will accelerate us towards building better neuromorphic computers and robots that are more natural for human interaction. Neuromorphic hardware can help connect models like Spaun to the world and to give these models a body”

 

Chris Eliasmith Photo Credit: uwaterloo.ca

Eliasmith and his team are accessing supercomputer Graham to program software specifically written to describe how each neuron functions within the brain model. Before Graham, it took Eliasmith 9,000 seconds to simulate one second of response. Today, SPAUN is running 10 times slower than real-time; a 900% improvement in performance with the help of Graham’s computational capacity.

“This increase in speed will accelerate us towards building better neuromorphic computers and robots that are more natural for human interaction. Neuromorphic hardware can help connect models like Spaun to the world and to give these models a body” says Eliasmith, who is currently working with several neuromorphic chips. Fusing these models with high performance computers is responsible for the myriad of innovations we hear about daily. Assisting big pharma by identifying effects of pharmaceutical drugs on cognitive behaviour and improving artificial intelligence by understanding the algorithms of the human brain, is how SPAUN is currently, and will continue to be, applied. The research conducted on SPAUN is the type of neural analysis behind innovations such as the autonomous vehicle.

To learn more about SPAUN and the work for Dr. Eliasmith, please visit the website for University of Waterloo’s Centre for Theoretical Neuroscience.