An algorithm that learns through rewards may show how our brain does too
By optimizing reinforcement-learning algorithms, DeepMind uncovered new details about how dopamine helps the brain learn. In 1951, Marvin Minsky, then …
Read more on technologyreview.com