arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2008.13607cs.LGstat.ML
收藏

排名政策决策

Ranking Policy Decisions

University of Oxford(牛津大学);University of Cambridge(剑桥大学);causaLens(causaLens公司);King’s College London(伦敦国王学院);Queen’s University Belfast(贝尔法斯特女王大学);Amazon(亚马逊)

Hadrien Pouget, Hana Chockler, Youcheng Sun, Daniel Kroening

摘要:

通过强化学习(RL)训练的策略通常是不必要的复杂,使得它们难以分析和解释。在使用$n$时间步长的运行中,策略将对要采取的操作做出$n$决策;我们推测,这些决策中只有一小部分比选择简单的默认操作更有价值。给定一个经过训练的策略,我们提出了一种基于统计故障定位的新的黑盒方法,该方法根据在这些状态中做出的决策的重要性对环境状态进行排名。我们认为,除其他外,国家的排名列表可以帮助解释和理解的政策。由于排名方法是统计的,很难对其质量进行直接评估。作为质量的代理,我们使用排名来创建新的,更简单的政策,从原来的修剪确定为不重要的决定(即,替换为默认操作),并衡量对性能的影响。我们在一组不同的标准基准测试上的实验表明,修剪后的策略可以在与原始策略相当的水平上执行。相反,我们表明,天真的方法排名的政策决定,例如,基于访问状态的频率的排名不会导致高性能的修剪策略。

英文摘要:

Policies trained via Reinforcement Learning (RL) are often needlessly complex, making them difficult to analyse and interpret. In a run with $n$ time steps, a policy will make $n$ decisions on actions to take; we conjecture that only a small subset of these decisions delivers value over selecting a simple default action. Given a trained policy, we propose a novel black-box method based on statistical fault localisation that ranks the states of the environment according to the importance of decisions made in those states. We argue that among other things, the ranked list of states can help explain and understand the policy. As the ranking method is statistical, a direct evaluation of its quality is hard. As a proxy for quality, we use the ranking to create new, simpler policies from the original ones by pruning decisions identified as unimportant (that is, replacing them by default actions) and measuring the impact on performance. Our experiments on a diverse set of standard benchmarks demonstrate that pruned policies can perform on a level comparable to the original policies. Conversely, we show that naive approaches for ranking policy decisions, e.g., ranking based on the frequency of visiting a state, do not result in high-performing pruned policies.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑