Technology company IBM is developing a new supercomputer powered by 64 TrueNorth chips for use by the US Air Force Research Lab, as part of a collaboration announced in June.
The new ‘brain-like’ supercomputer will be used for analytics involving pattern and object recognition, as well as ‘sensory processing’.
The TrueNorth neurosynaptic system has been designed to imitate human neurons and can convert audio, video and other forms of data received by sensors into symbols, which the computer can then process.
It is capable of performing advanced computations using less energy than conventional chips.
Air Force Research Laboratory principal electronics engineer Qing Wu said: “This is about building more intelligent machines that will work with humans to make human operators and analysts more effective and efficient when dealing with data.
“The major advantage of this chip is it runs machine-learning algorithms, the same ones as we run, the same functionality, same accuracy, but with much less power dissipation.”
The chip features a million neurons and was developed under the US Defense Advanced Research Projects Agency’s (DARPA) systems of neuromorphic adaptive plastic scalable electronics (SyNAPSE) programme.
“This is about building more intelligent machines that will work with humans to make human operators and analysts more effective and efficient when dealing with data.”
IBM brain-inspired computing group lead researcher Dharmendra Modha said that the neurons communicate throughout the system using patterns of pulses similar to the way human neurons use electrochemical pulses.
These are packed in clusters inside interconnected cores across the chip, with each core also holding components for information storage, processing, and communication.
TrueNorth capabilities will allow defence personnel to make better decisions by quickly combing through data and locating vital information.
The SyNAPSE programme seeks to develop computer systems with brain-like learning and problem-solving abilities.