This movie is a demonstration of learning of Multimodal Joint Attention with a infant-like robot through interactions with human.
And this movie is a later state of the learning.
After finishing of the learning, the robot could understand where or what human look at by observing human's faces or utterances.
Therefore, it can always look at an object which human attend to.
So, it knows what "bat" and "vehicle" are. Great. Now they just gotta teach him about "kill" and "humans" and he'll be all set!
MarcusKyo 10 months ago