arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1802.03501cs.AIcs.LGstat.ML
收藏

Tsallis熵正则化MDP中的路径一致性学习

Path Consistency Learning in Tsallis Entropy Regularized MDPs

Google Brain(谷歌大脑);Google DeepMind(谷歌DeepMind)

Ofir Nachum, Yinlam Chow, Mohammad Ghavamzadeh

摘要:

研究了稀疏熵正则化强化学习(ERL)问题,其中熵项是Tsallis熵的一种特殊形式.该公式的最优策略是稀疏的,即,~在每个状态下,只有少数动作具有非零概率。这解决了标准香农熵正则化RL(软ERL)公式的主要缺点,其中最佳策略是softmax,因此可能会将不可忽略的概率质量分配给非最佳动作。随着行动数量的增加,这一问题更加严重。在本文中,我们遵循Nachum等人的工作。(2017)在软ERL设置中,提出了一类新的路径一致性学习(PCL)算法,称为{ \em sparse PCL},用于稀疏ERL问题,可以使用策略和策略外数据。我们首先推导出一个{ \em稀疏一致性}方程,该方程指定了沿任意系统轨迹的稀疏ERL的最优价值函数和策略之间的关系。至关重要的是,一个弱形式的逆也是真的,我们量化的次优的政策,满足稀疏一致性,并表明,随着我们的行动的数量增加,这种次优是优于软ERL最优的政策。然后,我们使用这个结果来推导稀疏PCL算法。我们经验比较稀疏PCL与其软对应,并显示其优势,特别是在大量的行动的问题。

英文摘要:

We study the sparse entropy-regularized reinforcement learning (ERL) problem in which the entropy term is a special form of the Tsallis entropy. The optimal policy of this formulation is sparse, i.e.,~at each state, it has non-zero probability for only a small number of actions. This addresses the main drawback of the standard Shannon entropy-regularized RL (soft ERL) formulation, in which the optimal policy is softmax, and thus, may assign a non-negligible probability mass to non-optimal actions. This problem is aggravated as the number of actions is increased. In this paper, we follow the work of Nachum et al. (2017) in the soft ERL setting, and propose a class of novel path consistency learning (PCL) algorithms, called {\em sparse PCL}, for the sparse ERL problem that can work with both on-policy and off-policy data. We first derive a {\em sparse consistency} equation that specifies a relationship between the optimal value function and policy of the sparse ERL along any system trajectory. Crucially, a weak form of the converse is also true, and we quantify the sub-optimality of a policy which satisfies sparse consistency, and show that as we increase the number of actions, this sub-optimality is better than that of the soft ERL optimal policy. We then use this result to derive the sparse PCL algorithms. We empirically compare sparse PCL with its soft counterpart, and show its advantage, especially in problems with a large number of actions.

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。

↑