arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2010.01192cs.LGcs.AIcs.MA
收藏

多智能体通信中经验重放的修正

Correcting Experience Replay for Multi-Agent Communication

Gatsby Computational Neuroscience Unit(盖茨比计算神经科学单元);University College London(伦敦大学学院);Max Planck Institute for Biological Cybernetics(马克斯·普朗克生物控制论研究所);University of Tübingen(蒂宾根大学)

Sanjeevan Ahilan, Peter Dayan

摘要:

我们考虑使用多智能体强化学习(MARL)学习沟通的问题。一种常见的方法是使用从重放缓冲区采样的数据来学习离线策略。然而,过去接收的消息可能无法准确反映每个代理的当前通信策略,这使得学习变得复杂。因此,我们引入了一个“通信校正”,它占多智能体学习引起的观察到的通信的非平稳性。它的工作原理是重新标记接收到的消息,使其可能在通信者的当前策略下,从而更好地反映接收者的当前环境。为了说明代理商都是代理商和接收器的情况下,我们引入了一个有序的重新标记计划。我们的校正是计算效率高,可以与一系列的关闭政策的算法。我们在实验中发现,它大大提高了MARL系统在各种合作和竞争任务中学习的能力。

英文摘要:

We consider the problem of learning to communicate using multi-agent reinforcement learning (MARL). A common approach is to learn off-policy, using data sampled from a replay buffer. However, messages received in the past may not accurately reflect the current communication policy of each agent, and this complicates learning. We therefore introduce a 'communication correction' which accounts for the non-stationarity of observed communication induced by multi-agent learning. It works by relabelling the received message to make it likely under the communicator's current policy, and thus be a better reflection of the receiver's current environment. To account for cases in which agents are both senders and receivers, we introduce an ordered relabelling scheme. Our correction is computationally efficient and can be integrated with a range of off-policy algorithms. We find in our experiments that it substantially improves the ability of communicating MARL systems to learn across a variety of cooperative and competitive tasks.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑