ETH Zurich Proposes a Robotic System Capable of Self-Improving Its Semantic Perception | Synced

A research team from ETH Zurich combines continual learning and self-supervision to propose a novel robot system that enables online life-long self-supervised learning of semantic scene understanding.