DeepTH: Chip Placement with Deep Reinforcement Learning Using a Three-Head Policy Network | IEEE Conference Publication | IEEE Xplore