One of the biggest hurdles with teaching robots new skills is how to convert complex, high-dimensional data, like images from onboard RGB cameras, into actions that accomplish specific goals. Existing

Read more on