Self-Supervised Curriculum Generation for Autonomous Reinforcement Learning without Task-Specific Knowledge

arXiv:2311.09195v2 Announce Type: replace
Abstract: A significant bottleneck in applying current reinforcement learning algorithms to real-world scenarios is the need to reset the environment between every episode. This reset process demands substantial human intervention, making it difficult for the agent to learn continuously and autonomously. Several recent works have introduced autonomous reinforcement learning (ARL) algorithms that generate curricula for jointly training reset and forward policies. While their curricula can reduce the number of required manual resets by taking into account …

abstract agent algorithms arxiv autonomous cs.lg cs.ro current curriculum environment every human human intervention knowledge learn making process reinforcement reinforcement learning the environment type world