Cognitive computing has taken a giant step forward thanks to a new neurosynaptic chip from IBM that is designed to emulate the neurons and synapses in the human brain.
The architecture breaks away from the prevailing von Neumann model used for nearly 70 years. The CMOS chip is equipped with 5.4 billion transistors and can simulate 1 million neurons, 256 million synapses and 46 synaptic operations per second per watt. It consumes 70 milliwatts (mW), much less than the average microprocessor.
"These brain-inspired chips could transform mobility, via sensory and intelligent applications that can fit in the palm of your hand but without the need for Wi-Fi," said Dharmendra Modha, IBM fellow and chief scientist, brain-inspired computing, at IBM Research. He expects brain-inspired chips will lead to the creation of new generations of information technology systems, which will complement today's von Neumann machines.
The second-generation chip has taken nearly a decade to develop and was fabricated using Samsung's 28 nanometer process technology. IBM announced the initial single core hardware prototype for the chip three years ago and a software ecosystem for the chip one year ago.
To demonstrate scalability, IBM this month also revealed a 16-chip system with 16 million programmable neurons and 4 billion programmable synapses. IBM described its long-term goal as building a chip system with 10 billion neurons and a hundred trillion synapses, while consuming merely one kilowatt of power and occupying less than two liters of volume.
IBM and its collaborators, Cornell University and iniLabs, were awarded some $12 million in funding during 2013 from the Defense Advanced Research Projects Agency (DARPA) for Phase 3 of the Systems of Neuromorphic Adaptive Plastic Scalable Electronics (SyNAPSE) project. The cumulative funding awarded for the project since 2008 is around $53 million.
Synthesized cognition via neuromorphic electronic systems could be used in all manner of devices, including smartphones, that would benefit from intelligent sensor networks mimicking the brain's abilities for perception, action and cognition. They would be able to collect and assess raw data reflecting activities related to commerce, social, logistics, location, movement and environmental conditions.
Going forward, IBM said it is working on "integrating multisensory neurosynaptic processing into mobile devices constrained by power, volume and speed; integrating novel event-driven sensors with the chip; real-time multimedia cloud services accelerated by neurosynaptic systems; and neurosynaptic supercomputers by tiling multiple chips on a board, creating systems that would eventually scale to one hundred trillion synapses and beyond."
Numerous other companies are involved in efforts involving neuromorphic computing. For example, Qualcomm's (NASDAQ:QCOM) Zeroth neuromorphic chip program has been focused on signing up researchers this year to test its technology as the company eyes the commercial introduction of brain-mimicking chips as soon as 2015.
- see this IBM release
- see this EE Times article
Baidu's Silicon Valley R&D center targets deep learning
Qualcomm demos robot with neuromorphic chip
TeraDeep hopes to see Apple, Qualcomm, others use its artificial intelligence technology
Qualcomm, Ericsson and AT&T execs tout wearables, neuromorphic computing and more
Brain chips offer a new approach to crunching data