arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2207.04136cs.LGcs.AIcs.RO
收藏

CompoSuite:一个组合式强化学习基准

CompoSuite: A Compositional Reinforcement Learning Benchmark

University of Pennsylvania(宾夕法尼亚大学)

Jorge A. Mendez, Marcel Hussing, Meghna Gummadi, Eric Eaton

摘要:

我们提出了CompoSuite,一个开源的模拟机器人操作基准组合多任务强化学习(RL)。每个CompoSuite任务都需要一个特定的机器人手臂来操纵一个单独的物体,以实现任务目标,同时避免障碍物。这种任务的组合定义赋予了CompoSuite两个显著的属性。首先,改变机器人/物体/目标/障碍物元素会导致数百个RL任务,每个任务都需要有意义的不同行为。第二,RL方法可以专门评估其学习任务组成结构的能力。后一种功能分解问题的能力将使智能代理能够识别和利用学习任务之间的共性,以处理各种各样的高度多样化的问题。我们对现有的单任务、多任务和组合学习算法在各种训练环境下进行了基准测试,并评估了它们在组合上推广到看不见的任务的能力。我们的评估暴露了现有的RL方法的缺点方面的组合性,并开辟了新的途径进行调查。

英文摘要:

We present CompoSuite, an open-source simulated robotic manipulation benchmark for compositional multi-task reinforcement learning (RL). Each CompoSuite task requires a particular robot arm to manipulate one individual object to achieve a task objective while avoiding an obstacle. This compositional definition of the tasks endows CompoSuite with two remarkable properties. First, varying the robot/object/objective/obstacle elements leads to hundreds of RL tasks, each of which requires a meaningfully different behavior. Second, RL approaches can be evaluated specifically for their ability to learn the compositional structure of the tasks. This latter capability to functionally decompose problems would enable intelligent agents to identify and exploit commonalities between learning tasks to handle large varieties of highly diverse problems. We benchmark existing single-task, multi-task, and compositional learning algorithms on various training settings, and assess their capability to compositionally generalize to unseen tasks. Our evaluation exposes the shortcomings of existing RL approaches with respect to compositionality and opens new avenues for investigation.

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。

↑