What's harder, perception or control for these problems? So being able to perfectly perceive everything or figuring out a plan once you perceive everything, how to interact with all the agents in the environment. In your sense, from a learning perspective, is perception or action harder in that giant, beautiful, multi-task learning neural network?

The hardest thing is having accurate representation of the physical objects in vector space. So taking the visual input, primarily visual input, some sonar and radar, and then creating an accurate vector space representation of the objects around you. Once you have an accurate vector space representation, the planning and control is relatively easier.

That is relatively easy. Basically once you have accurate vector space representation, then you're kind of like a video game. Cars in Grand Theft Auto or something, they work pretty well. They drive down the road, they don't crash, pretty much, unless you crash into them. That's because they've got an accurate vector space representation of where the cars are, and then rendering that as the output.

Page 1 of 9 Page 2 of 9 Page 3 of 9 Page 4 of 9 Page 5 of 9 Page 6 of 9 Page 7 of 9 Page 8 of 9 Page 9 of 9 Page 10 of 10 Page 11 of 10 Page 12 of 10 Page 13 of 10 Page 14 of 10 Page 15 of 15 Page 16 of 15 Page 17 of 15 Page 18 of 15 Page 19 of 15 Page 20 of 20