arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2201.10081cs.LGcs.AI
收藏

动态感知的学习奖励函数比较

Dynamics-Aware Comparison of Learned Reward Functions

Toyota Research Institute (TRI)(丰田研究院(TRI));University of California, Berkeley(加州大学伯克利分校)

Blake Wulfe, Ashwin Balakrishna, Logan Ellis, Jean Mercat, Rowan McAllister, Adrien Gaidon

摘要:

学习奖励函数的能力在现实世界中部署智能代理方面发挥着重要作用。然而,比较奖励函数,例如作为评估奖励学习方法的一种手段,提出了一个挑战。奖励函数通常通过考虑优化策略的行为来进行比较,但这种方法将奖励函数中的缺陷与用于优化它的策略搜索算法的缺陷混为一谈。为了解决这一挑战,Gleave et al.(2020)提出了等价策略不变比较(EPIC)距离。EPIC避免了策略优化,但这样做需要在转换时计算奖励值,这在系统动态下可能是不可能的。这对于学习的奖励函数来说是有问题的,因为它需要在训练分布之外对它们进行评估,导致不准确的奖励值,我们发现这会使EPIC在比较奖励时无效。为了解决这个问题,我们提出了动态感知奖励距离(DARD),一个新的奖励伪度量。DARD使用环境的近似转换模型将奖励函数转换为允许比较的形式,该比较对于奖励塑造是不变的,同时仅在接近其训练分布的转换上评估奖励函数。在模拟物理域的实验表明,DARD使可靠的奖励比较没有政策优化,是显着更具有预测性比基线方法的下游政策性能时,处理学习的奖励功能。

英文摘要:

The ability to learn reward functions plays an important role in enabling the deployment of intelligent agents in the real world. However, comparing reward functions, for example as a means of evaluating reward learning methods, presents a challenge. Reward functions are typically compared by considering the behavior of optimized policies, but this approach conflates deficiencies in the reward function with those of the policy search algorithm used to optimize it. To address this challenge, Gleave et al. (2020) propose the Equivalent-Policy Invariant Comparison (EPIC) distance. EPIC avoids policy optimization, but in doing so requires computing reward values at transitions that may be impossible under the system dynamics. This is problematic for learned reward functions because it entails evaluating them outside of their training distribution, resulting in inaccurate reward values that we show can render EPIC ineffective at comparing rewards. To address this problem, we propose the Dynamics-Aware Reward Distance (DARD), a new reward pseudometric. DARD uses an approximate transition model of the environment to transform reward functions into a form that allows for comparisons that are invariant to reward shaping while only evaluating reward functions on transitions close to their training distribution. Experiments in simulated physical domains demonstrate that DARD enables reliable reward comparisons without policy optimization and is significantly more predictive than baseline methods of downstream policy performance when dealing with learned reward functions.

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。

↑