New Computer Chip Mimics Brain Function
IBM has developed a radically different computer chip whose architecture seeks to mimic that of the brain. A prototype was first built in 2011 based on designs from a monkey brain. It simulates the functions of neurons and synapses. The chip took 10 years to develop and relied on $53.5 million in funding from the Defense Advanced Research Projects Agency (DARPA). Unlike conventional chips, this new one excels at tasks such as recognizing patterns and classifying objects
Dubbed, TrueNorth, the chip uses 5.4 billion transistors, but draws just 70 milliwatts of power, about the same as a hearing aid. This compares to today’s conventional PC processors that have about 1.4 billion transistors and draw many times more power, about 35 to 140 watts. The chip was built for IBM by Samsung using standard digital technology like that it uses to make microprocessors for smartphones. Other efforts at building new types of chips have often focused on the use of esoteric materials or production processes
TrueNorth chips can be linked to build large, complex systems such as room-size supercomputers, but IBM also hopes to offer smaller, simpler chips for applications where saving space is critical. The company sees great potential for this new type of chip and is already talking to potential partners about collaborating to bring the chip to market.
Dharmendra Modha, a researcher at IBM’s Almaden Research Center writes, “The architecture can solve a wide class of problems from vision, audition, and multi-sensory fusion, and has the potential to revolutionize the computer industry by integrating brain-like capability into devices where computation is constrained by power and speed. These systems can efficiently process high-dimensional, noisy sensory data in real time, while consuming orders of magnitude less power than conventional computer architectures.”
Others are not so optimistic about the outlook for TrueNorth-type chips. An article in the Boston Globe quotes an email sent by Yann Lecun the director of artificial intelligence research at Facebook in which he says, “The chip appears to be very limited in many ways, and the performance is not what it seems.