The new model could allow robots to grasp concepts like gravity and object permanence while relying less on large amounts of video or training data.
The new model could allow robots to grasp concepts like gravity and object permanence while relying less on large amounts of video or training data.