Right and Left Brains Collide
Military and industry researchers combine best of both worlds.
The U.S. Air Force Research Laboratory (AFRL) and IBM are collaborating on a brain-inspired supercomputing system powered by a 64-chip array. The laboratory is investigating applications for the system in embedded, mobile, autonomous settings where limiting factors today include size, weight and power.
As an end-to-end software ecosystem, the scalable platform would enable deep neural-network learning and information discovery. Its advanced pattern recognition and sensory processing power would be the equivalent of 64 million neurons and 16 billion synapses; however, the processor component only will consume approximately 10 watts, the equivalent of a dim light bulb.
The 64-chip array is part of the company’s TrueNorth Neurosynaptic System, which can convert data such as images, video, audio and text from multiple, distributed sensors into symbols in real time. IBM researchers believe the system’s design will be more efficient for pattern recognition and integrated sensory processing than systems using conventional chips.
AFRL researchers will combine this right-brain perception capability system with the left-brain symbol processing capabilities of conventional computer systems. The large scale of the system will enable both data parallelism and model parallelism. As a result, multiple data sources can be run in parallel against the same neural network and independent neural networks can be run in parallel on the same data.
“AFRL was the earliest adopter of TrueNorth for converting data into decisions,” says Daniel S. Goddard, director of the lab's information directorate. “The new neurosynaptic system will be used to enable new computing capabilities important to AFRL’s mission to explore, prototype and demonstrate high-impact, game-changing technologies that enable the Air Force and the nation to maintain its superior technical advantage.”
The IBM TrueNorth Neurosynaptic System was originally developed under the auspices of the Defense Advanced Research Projects Agency’s Systems of Neuromorphic Adaptive Plastic Scalable Electronics program in collaboration with Cornell University.