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 2205.07802cs.LGcs.AIstat.ML
收藏

深度强化学习中的优先偏差

The Primacy Bias in Deep Reinforcement Learning

Mila(米拉人工智能研究所);Université de Montréal(蒙特利尔大学)

Evgenii Nikishin, Max Schwarzer, Pierluca D'Oro, Pierre-Luc Bacon, Aaron Courville

摘要:

这项工作确定了深度强化学习(RL)算法的一个常见缺陷:倾向于依赖早期的交互,而忽略后来遇到的有用证据。由于在不断增长的数据集上进行训练,深度RL代理会产生过度拟合早期经验的风险,从而对学习过程的其余部分产生负面影响。受认知科学的启发,我们将这种效应称为首因偏见。通过一系列实验,我们剖析了加剧这种偏见的深度RL的算法方面。然后,我们提出了一个简单而普遍适用的机制,处理的首要偏见定期重置代理的一部分。我们将这种机制应用于离散(Atari 100k)和连续动作(DeepMind Control Suite)领域的算法,不断提高它们的性能。

英文摘要:

This work identifies a common flaw of deep reinforcement learning (RL) algorithms: a tendency to rely on early interactions and ignore useful evidence encountered later. Because of training on progressively growing datasets, deep RL agents incur a risk of overfitting to earlier experiences, negatively affecting the rest of the learning process. Inspired by cognitive science, we refer to this effect as the primacy bias. Through a series of experiments, we dissect the algorithmic aspects of deep RL that exacerbate this bias. We then propose a simple yet generally-applicable mechanism that tackles the primacy bias by periodically resetting a part of the agent. We apply this mechanism to algorithms in both discrete (Atari 100k) and continuous action (DeepMind Control Suite) domains, consistently improving their performance.

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。

↑