Researchers are looking to biology—specifically neuroscience—for ways to meet accelerating demands on computer capacity. Read about their work in a New York Times article by Steve Lohr titled “Creating Artificial Intelligence Based on the Real Thing.”
A team of researchers is at work developing “neuromorphic chips,” which are based on how the brain functions (ie, via neurons and synapses). The goal is create computers that learn from experience without consuming vast quantities of power. The human brain runs on 20 watts of power; IBM’s Watson—the computer that won that game of Jeopardy—runs on 85,000 watts.
