Samsung wants to copy the structure of the human brain onto computer chips

Samsung wants to copy the structure of the human brain onto computer chips

28. September 2021 0 Von Horst Buchwald

Samsung wants to copy the structure of the human brain onto computer chips

Seoul, Sept. 28, 2021

The Korean megacorporation has teamed up with Harvard University to replicate the structure of the brain in a chip format, hoping this will give future chips access to the brain’s capabilities of „low power consumption, easy learning, adaptation to the environment, and even autonomy and cognition.“
Samsung knows how this works, „The brain is made up of a large number of neurons whose wiring map is responsible for the brain’s functions. So knowing this map is the key to reverse engineering the brain.“
The team consists of the following four people: Harvard professors Donhee Ham and Hongkun Park, Samsung SDS president and CEO Sungwoo Hwang, and Samsung Electronics VC and CEO Kinam Kim.
Their „perspective paper“ – „Neuromorphic electronics based on copying and pasting the brain,“ published in Nature Electronics – suggests that by copying the brain’s complex mechanics and transferring them to Samsung’s memory technology, electronics may be able to mimic some of the brain’s properties.

https://www.nature.com/articles/s41928-021-00646-1

The core of this vision is to „copy“ the functional synaptic connectivity map of a mammalian neural network using advanced neuroscience tools and then „paste“ that map onto a high-density three-dimensional network of solid-state memories. Our copy-and-paste approach could potentially lead to silicon integrated circuits that better match the brain’s computational properties,“ the paper states.
The team plans to accomplish this by „using a breakthrough nanoelectrode array developed by Dr. Ham and Dr. Park and inserting this map into a three-dimensional network of high-density solid-state memories.“
It’s a task whose difficulty should not be underestimated, because, „Since the human brain is estimated to have about 100 billion neurons and about a thousand times more synaptic connections, the ultimate neuromorphic chip will require about 100 trillion memories.“
Since this is a „perspective paper,“ the reader also learns nothing about when the researchers will reach their goal.