Reinforcement Learning Guided by Provable Normative Compliance
This work addresses the challenge of ensuring normative compliance in autonomous agents, but it is incremental as it builds on an existing framework and focuses on generalization rather than a breakthrough.
The paper tackles the problem of generalizing punishment assignment for safe and ethical behavior in reinforcement learning agents by using a normative supervisor to dynamically translate states into defeasible deontic logic theories and applying multi-objective RL to balance ethical and non-ethical objectives, showing effectiveness regardless of punishment magnitude.
Reinforcement learning (RL) has shown promise as a tool for engineering safe, ethical, or legal behaviour in autonomous agents. Its use typically relies on assigning punishments to state-action pairs that constitute unsafe or unethical choices. Despite this assignment being a crucial step in this approach, however, there has been limited discussion on generalizing the process of selecting punishments and deciding where to apply them. In this paper, we adopt an approach that leverages an existing framework -- the normative supervisor of (Neufeld et al., 2021) -- during training. This normative supervisor is used to dynamically translate states and the applicable normative system into defeasible deontic logic theories, feed these theories to a theorem prover, and use the conclusions derived to decide whether or not to assign a punishment to the agent. We use multi-objective RL (MORL) to balance the ethical objective of avoiding violations with a non-ethical objective; we will demonstrate that our approach works for a multiplicity of MORL techniques, and show that it is effective regardless of the magnitude of the punishment we assign.