Cortex-inspired learning key to AI, says expertUnsupervised learning is one of the keys to achieving human-level mental abilities in artificial intelligence (AI), Sebastian Seung, one of the leading scientists in the field, said Thursday.
“Cortex-inspired unsupervised learning, or also called self-supervised learning, is key to achieving human-level AI,” Seung, a professor at the Princeton Neuroscience Institute and Department of Computer Science, said during his keynote speech at the Samsung AI Forum 2018 held in southern Seoul.
The annual forum - hosted by Korea’s tech giant Samsung Electronics - is an academic event for the foremost minds in the world to come together and share their views on the current and future development of AI technology.
Seung has done influential research in both computer science and neuroscience. He was recently chosen to oversee machine learning projects at Samsung Electronics’ research division.
Seung emphasized the importance of the cortex - the outer surface of the cerebrum, composed of gray matter - in solving the mystery behind human brains and in mapping the brain’s wiring.
“[The] cortex is the biggest structure of [the] human brain. [The] human cortex is much bigger than [that of] mice,” he stressed, citing a project he recently spearheaded to observe activity and connectivity among neurons.
Last year, Seung and his collaborators successfully used AI to reconstruct all the neural wiring within a 0.001 cubic millimeter chunk of mouse cortex, showing insights such as how the circuits are connected and how they perform computations.
Seung, who is known for his efforts to communicate neuroscience to the general public, created a website called EyeWire that has brought together 150,000 people from 130 countries to play a game to map neural connections.