SARSA(0) Reinforcement Learning over Fully Homomorphic Encryption

SARSA(0) Reinforcement Learning over Fully Homomorphic Encryption
复制标题

SARSA(0) 全同态加密的强化学习

DOI:
--
复制
发表时间:
2020
期刊:
2021 SICE International Symposium on Control Systems (SICE ISCS)
影响因子:
--
通讯作者:
Takashi Tanaka
Takashi Tanaka
中科院分区:
--
文献类型:
--
作者:
Jihoon Suh;Takashi Tanaka

文献摘要

被引文献

相似文献

我们考虑了一个基于云的控制架构,其中本地工厂将控制合成任务外包给云。特别是,我们考虑基于云的强化学习(RL),其中更新值函数被外包给云。为了实现机密性,我们实现了全同态加密(FHE)的计算。我们使用一个CKKS加密方案和一个修改的SARSA(0)强化学习,以纳入预防引起的延迟。然后,我们给出了SARSA(0)的延迟更新规则与阻塞机制的收敛性结果。最后,我们通过一个经典的极点平衡问题的实现给出了一个数值演示。
We consider a cloud-based control architecture in which the local plants outsource the control synthesis task to the cloud. In particular, we consider a cloud-based reinforcement learning (RL), where updating the value function is outsourced to the cloud. To achieve confidentiality, we implement computations over Fully Homomorphic Encryption (FHE). We use a CKKS encryption scheme and a modified SARSA(0) reinforcement learning to incorporate the encryption-induced delays. We then give a convergence result for the delayed updated rule of SARSA(0) with a blocking mechanism. We finally present a numerical demonstration via implementing on a classical pole-balancing problem.