Intel has introduced a set of 64 chips that mimic the human brain. The chipset is equivalent to 8 million neurons, it is composed of Loihi chips and has been called Pohoiki Beach.
Even though the IA it tries to imitate the behavior of our brain, the technology we have does not succeed. A greater need for collection, analysis and decision-making using natural, unstructured and highly dynamic data is driving demand for a new kind of computing.
Now it looks like we’re a little closer to improving AI, thanks to Intel. The company has presented a system that is capable of mimicking the human brainaccording to PCMag.
To achieve this they have used neuromorphic computing, an interdisciplinary engineering based on biology, physics, mathematics, computer science and electronic engineering to design artificial neural systems whose physical structure and design principles are based on nervous biological systems.
The system is composed of 64 chips equivalent to the calculation of 8 million human neurons, although our brain is composed of 80 billion neurons.
At the moment it will only be available to Intel partner researchers
Although it seems that the chip could not mimic a human brain by not reaching the same number of neurons, the chipset can calculate 1,000 times faster and 10,000 times more efficiently than the chips we currently have.
The chipset is called Pohoiki Beach and it’s based on Loihi, a chip that Intel introduced last September 2017. Loihi is a single chip it mimics the functions of the brain, learning to function based on different modes of response to stimuli from the environment.
This energy-efficient chip uses data to learn and make inferences, gets smarter over time and it does not need to be trained in a traditional way, as they are based on the ways of communication and learning of neurons.
Pohoiki Beach he plans to use it in Artificial Intelligence, as he manages to adapt better to changes and uses the knowledge he already has to learn exponentially over time.
What makes the great Intel chip stand out is its energy consumption. Pohoiki Beach 100 times less power than a normal processor when performing Machine Learning calculations.