The brain can process high volumes of data faster than any single computer. It also uses a surprisingly low amount of power. A shortcut to replicating these features may come sooner than expected, thanks to “neuromorphic” chips envisioned by Samsung and Harvard researchers.
This is the fourth blue sky idea that Lex explores in today’s thematic column.
Such artificial intelligence on a human level will have wide applications. In medicine, for example, where it can provide diagnoses. The biggest challenge was to train machines to process information in a “brain-like” way. Computers are linear and methodical. Brains are lateral and make sudden jumps.
AI researchers try to circumvent this by feeding computers large amounts of data. Programs are just as good as this app. The limited quantity and quality of available data often sent human-like AI on the wrong path, prompting them to illogical expressions and offensive opinions.
It was also difficult for current AI models to make their own decisions. They struggle to prioritize data, such as the information needed to learn a certain function more effectively.
The brain, on the other hand, absorbs new concepts without explicit instruction and often comes to intuitive conclusions that reduce the need for background data.
Samsung, working with Harvard researchers, seeks to overcome these challenges a scheme for the “copy and paste” of the brain’s more than 100 billion neurons and their connections to new types of 3D “neuromorphic” chips. It is supposed to repeat the brain’s complex interconnections of neurons and synapses, to overcome the limitations of conventional microprocessors.
But allocating chip manufacturing production to support this blue sky research can be difficult amid shortages of normal semiconductors. In addition, advanced AI is fraught with ethical and legal issues. It can be difficult for Samsung – a company that focuses on engineering and manufacturing – to navigate.
However, the swing factor will be humanity’s unlimited appetite for data processing capacity. The replication of human thinking will remain difficult as scientists struggle to understand it. But chips that came close to doing so would have a huge potential market.
This is the fourth article on blue sky thinking published by Lex today. Look out for the others in Lex online.