提出让智能体可关闭的新训练方法，避免其抗拒停机。
Shutdownable Agents through POST-Agency
- 训练智能体只在相同长度轨迹间比较偏好
- 理论证明能确保智能体不关心任务时长分布
- 适合希望控制智能体、又需其高效执行的开发者
许多人担心未来的人工智能代理会抵抗关闭。本文提出一种名为 POST-Agents 的方案，即训练智能体仅在相同长度的轨迹之间进行偏好判断（POST）。随后证明，在其他条件下，POST 可推出 Neutrality+：智能体最大化期望效用，且忽略轨迹长度的概率分布。本文认为，Neutrality+ 能保证智能体可被关闭，同时仍保持有用性。
原文摘要 · Abstract (English)
Many fear that future artificial agents will resist shutdown. I present an idea - the POST-Agents Proposal - for ensuring that doesn't happen. I propose that we train agents to satisfy Preferences Only Between Same-Length Trajectories (POST). I then prove that POST - together with other conditions - implies Neutrality+: the agent maximizes expected utility, ignoring the probability distribution over trajectory-lengths. I argue that Neutrality+ keeps agents shutdownable and allows them to be useful.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。