GIDSeg: Learning 3D Reinforcement Learning from Sparse annotations - 42Papers