Abstract: Artificial intelligence systems are associated with inherent risks, such as uncontrollability and lack of interpretability. To address these risks, we need to develop artificial intelligence ...