arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2107.06405cs.LGcs.AIcs.RO
收藏

稀疏奖励任务的最短路径约束强化学习

Shortest-Path Constrained Reinforcement Learning for Sparse Reward Tasks

University of Michigan(密歇根大学);LG AI Research(LG人工智能研究院);Yonsei University(延世大学);Microsoft Research(微软研究院)

Sungryull Sohn, Sungtae Lee, Jongwook Choi, Harm van Seijen, Mehdi Fatemi, Honglak Lee

摘要:

我们提出了k-最短路径(k-SP)约束:一种新的约束代理的轨迹,提高了样本效率稀疏奖励MDP。我们证明了任何最优策略必然满足k-SP约束。值得注意的是,k-SP约束防止策略沿着非k-SP轨迹探索状态-动作对(例如,来回的)。然而,在实践中,排除状态动作对可能会阻碍RL算法的收敛。为了克服这一点,我们提出了一种新的成本函数,惩罚违反SP约束的政策,而不是完全排除它。我们的数值实验在一个表格RL设置表明,SP约束可以显着减少策略的轨迹空间。因此,我们的约束通过抑制冗余的探索和利用,使更多的样本有效的学习。我们在MiniGrid、DeepMind Lab、Atari和Fetch上的实验表明,该方法显着提高了近端策略优化(PPO),即使在连续控制任务中也优于现有的新颖性探索方法,包括基于计数的探索,这表明它通过防止代理采取冗余动作来提高样本效率。

英文摘要:

We propose the k-Shortest-Path (k-SP) constraint: a novel constraint on the agent's trajectory that improves the sample efficiency in sparse-reward MDPs. We show that any optimal policy necessarily satisfies the k-SP constraint. Notably, the k-SP constraint prevents the policy from exploring state-action pairs along the non-k-SP trajectories (e.g., going back and forth). However, in practice, excluding state-action pairs may hinder the convergence of RL algorithms. To overcome this, we propose a novel cost function that penalizes the policy violating SP constraint, instead of completely excluding it. Our numerical experiment in a tabular RL setting demonstrates that the SP constraint can significantly reduce the trajectory space of policy. As a result, our constraint enables more sample efficient learning by suppressing redundant exploration and exploitation. Our experiments on MiniGrid, DeepMind Lab, Atari, and Fetch show that the proposed method significantly improves proximal policy optimization (PPO) and outperforms existing novelty-seeking exploration methods including count-based exploration even in continuous control tasks, indicating that it improves the sample efficiency by preventing the agent from taking redundant 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。

↑