SUNUM researcher Kaan Demirel discussed the issue of excessive energy consumption—one of the most significant challenges facing large artificial intelligence systems today—in an interview with GazeteSU.
"Neuromorphic computing is a new computing approach inspired by the working principles of the human brain. While large artificial intelligence systems consume very high amounts of energy today, the human brain can perform extremely complex operations with only about 20 watts of power. Therefore, one of the most important goals in future artificial intelligence hardware is to transfer this energy efficiency of the brain to electronic systems.
In this field, I am working on next-generation electronic memory units that mimic the behavior of synapses in our brain using two-dimensional (2D) materials and thin-film technologies. These studies aim to contribute to the development of faster, lower-power consuming artificial intelligence systems with less environmental impact in the future."
Please click to read more: gazetesu.sabanciuniv.edu
