Samsung Electronics is developing a new chip that mimics the manner in which the human brain operates, seeking to deliver a solution for artificial intelligence platforms of the future. The tech giant enlisted the help of a number of Korean experts such as the KAIST public research institute and the Seoul National University, with the partners already working on turning their idea into a reality.
The concept of a chip designed to operate like a human brain is far from new and has been iterated and revised for many years now. At its core, Samsung’s chip is meant to work in much the same way like numerous solutions that preceded it, with its components being dedicated to simulating a nervous system. Coupled with an extreme level of energy efficiency that’s said to require a hundred-millionth of electricity used by contemporary chips, the device is meant to deliver a next-generation platform for a wide variety of AI use cases. Samsung is reportedly committing approximately $8.5 million to financing initial research required for the chip which is estimated to take three years, after which the company is expected to ramp up its investment, though no full timeframe for the project has yet been provided.
Simulating neurons is an expensive ambition and one that the tech industry has yet to bring anywhere close to the level capable of mimicking an actual human brain which has around 100 billion neural cells on average. Intel came the closest to doing so until now but even its high-end TrueNorth chip simulates “only” 260 million neurons and hence hasn’t even reached the processing capacity of many insects. It’s presently unclear what artificial neuron count is Samsung targeting with its supposedly revolutionary chip and whether the company is even able to surpass whatever the current industry leaders come up with several years from now. Potential use cases of the chip itself are practically endless, just like AI applications in general are; Samsung’s current commitment to the segment largely revolves around its smartphones and Internet of Things offerings but an energy-efficient AI platform would also be able to improve all of the company’s other products, services, and even manufacturing operations.