Reliability quantification of deep reinforcement learning (DRL)-based control is a significant challenge for the practical application of artificial intelligence (AI) in safety-critical systems. This study proposes a method for quantifying the reliability of DRL-based control. First, an existing method, random noise distillation, was applied to the reliability evaluation to clarify the issues to be solved. Second, a novel method for reliability quantification was proposed to solve these issues. The reliability is quantified using two neural networks: reference and evaluator. They have the same structure with the same initial parameters. The outputs of the two networks were the same before training. During training, the evaluator network parameters were updated to maximize the difference between the reference and evaluator networks for trained data. Thus, the reliability of the DRL-based control for a state can be evaluated based on the difference in output between the two networks. The proposed method was applied to DQN-based control as an example of a simple task, and its effectiveness was demonstrated. Finally, the proposed method was applied to the problem of switching trained models depending on the state. Con-sequently, the performance of the DRL-based control was improved by switching the trained models according to their reliability.

翻译：基于深度强化学习（DRL）的控制在实际应用于安全关键型人工智能（AI）系统时，其可靠性量化是一项重大挑战。本研究提出了一种量化DRL控制可靠性的方法。首先，将现有方法——随机噪声蒸馏——应用于可靠性评估，以明确待解决的问题。其次，针对这些问题提出了一种新的可靠性量化方法。该方法通过两个神经网络（参考网络与评估网络）进行可靠性量化：这两个网络具有相同的结构和初始参数，训练前输出一致。训练过程中，评估网络参数被更新，以最大化其在训练数据上相对于参考网络的输出差异。因此，基于两个网络输出之间的差异，可评估DRL控制对特定状态下的可靠性。将该方法应用于基于DQN的简单任务控制示例，验证了其有效性。最后，将所提方法应用于根据状态切换训练模型的问题。结果表明，通过根据可靠性切换训练模型，DRL控制的性能得到了提升。