自监督对抗模仿学习（英文版）

英文标题：Self-Supervised Adversarial Imitation Learning中文摘要：本文提出一个包含鉴别器的行为克隆学习方法，用于解决之前的学习策略容易被困入错误局部最小值的问题，避免了人工干预的需要，利用鉴别器计算得到过渡函数从而帮助学习。英文摘要：Behavioural cloning is an imitation learning technique that teaches an agenthow to behave via expert demonstrations. Recent approaches use self-supervisionof fully-observable unlabelled snapshots of the states to decode state pairsinto actions. However, the iterative learning scheme employed by thesetechniques is prone to get trapped into bad local minima. P