Computer-chip industry giant Intel has unveiled its self-learning chip – codenamed “Loihi” – that mimics how the human brain functions.
According to Dr. Michael Mayberry, corporate vice president and managing director of Intel Labs at Intel Corporation, Loihi’s digital circuits mimic the brain’s basic mechanics “from how neurons communicate and learn, using spikes and plastic synapses that can be modulated based on timing.” He said mimicking the brain’s basic mechanics “could help computers self-organize and make decisions based on patterns and associations.”
The potential benefits of self-learning chips are limitless, according to Mayberry. Self-learning chips, he said, can be used to read a person’s heartbeat at various conditions – after a meal, after jogging or before going to bed. The data gathered will be processed to find the “normal” heartbeat. If the system detects a pattern that does not match the normal pattern, it will then flag such incident.
Self-learning chips, according to Mayberry, can similarly be used in cyber security where a pattern that does not match the normal could be flagged by the system as this could be a sign of a hack or breach in the system.
The Loihi chip has a total of 130,000 neurons with each neuron capable of communicating with thousands of other neurons. Researchers at Intel demonstrated that the Loihi chip is 1,000 times more energy-efficient compared to general purpose computing required for typical training systems.
The researchers also demonstrated that the Loihi chip learns faster at 1 million times improvement compared “with other typical spiking neural nets as measured by total operations to achieve a given accuracy when solving MNIST digit recognition problems.”
The company plans to share the Loihi chip in the first half of 2018 with leading university and research institutions to accelerate artificial intelligence.