Intel has developed a new chip, code-named Loihi, that mimics how the brain works.
The human brain consists of a network of neurons that link together by forming synapses – for that matter all brains in the natural world work that way. The idea of applying that approach to computing is not new, neural networks also attempt to mimic the brain and IBM announced its own neural chip several years ago. But the Intel product certainly appears to take the concept a good deal further, applying what it calls neuromorphic chips and says its “model draws inspiration from how neurons communicate and learn, using spikes and plastic synapses that can be modulated based on timing.”
According to Intel’s Dr. Michael Mayberry, the chip is extremely energy efficient, “uses data to learn and make inferences, gets smarter over time and does not need to be trained in the traditional way.” The company claims that the Loihi chip can learn 1,000 times faster than other spiked neural nets.
Intel says that “the Loihi test chip offers highly flexible on-chip learning and combines training and inference on a single chip. This allows machines to be autonomous and to adapt in real time instead of waiting for the next update from the cloud.”
It’s exciting stuff, but put it in context. The Intel chip will have 130,000 neurons, the human brain has 100 billion. The Intel chip forms 130 million synapses, the human brain 100 trillion. Then again, the human brain evolved from far simpler brains, and in the digital sphere, evolution can work several orders of magnitude faster.
Dr. Mayberry said: “In a future with neuromorphic computing, all of what you can imagine – and more – moves from possibility to reality, as the flow of intelligence and decision-making becomes more fluid and accelerated.”