Knowledge Transfer for Deep Reinforcement Learning with Hierarchical Experience Replay

• Main Authors: Yin, Haiyan, Pan, Sinno Jialin
• 其他作者: School of Computer Science and Engineering
• 格式: Conference or Workshop Item
• 語言: English
• 出版: 2017
• 主題: Deep reinforcement learning; Policy distillation
• 在線閱讀: https://hdl.handle.net/10356/83043; http://hdl.handle.net/10220/42453; https://aaai.org/ocs/index.php/AAAI/AAAI17/paper/view/14478

The process for transferring knowledge of multiple reinforcement learning policies into a single multi-task policy via distillation technique is known as policy distillation. When policy distillation is under a deep reinforcement learning setting, due to the giant parameter size and the huge state s...