Abstract: Traditional deep reinforcement learning (DRL) exhibits training inefficiencies in complex robot assembly tasks. Behavior Cloning (BC) can alleviate this issue by providing expert guidance.