Samsung Co-CEO Han Jong-hee Dies at 63 from Cardiac Arrest

Researchers at the Korea Advanced Institute of Science and Technology (KAIST) have developed a self-learning memristor that could enhance neuromorphic computing and aid in replicating human brain functions. This chip, which can learn from errors, may allow artificial intelligence (AI) systems to perform tasks locally with improved energy efficiency and privacy, as revealed in a study published in Nature Electronics.
Historically conceptualized by American engineer Leon Chua in 1971, memristors function as "memory resistors" and simulate synaptic activity in the brain. KAIST's recent advancement enables the chip to separate moving images from backgrounds and become more proficient over time. Researchers Hakcheon Jeong and Seungjae Han emphasized that this technology mimics the brain's efficient processing, likening it to a streamlined workspace.
In tandem with its memristor innovations, KAIST has also unveiled its first AI superconductor chip, capable of high-speed processing while consuming minimal power. While these developments inch closer to creating a brain-on-a-chip, experts caution that such advancements do not equate to machines achieving comprehensive human-like intelligence.