arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2310.09615cs.LG
收藏

STORM:用于强化学习的高效随机Transformer世界模型

STORM: Efficient Stochastic Transformer based World Models for Reinforcement Learning

Beijing Institute of Technology(北京理工大学);Beijing Institute of Technology Chongqing Innovation Center(北京理工大学重庆创新中心);Tsinghua University(清华大学)

Weipu Zhang, Gang Wang, Jian Sun, Yetian Yuan, Gao Huang

AI总结:

STORM结合Transformer与变分自编码器构建高效随机世界模型,在Atari 100k基准上达到126.7%平均人类性能,训练时间仅需4.3小时。

摘要:

近年来,基于模型的强化学习算法在视觉输入环境中已展现出显著的效果。这些方法首先通过自监督学习构建真实环境的参数化仿真世界模型。通过利用世界模型的想象,智能体的策略得以增强,无需受限于从真实环境中采样的约束。这些算法的性能在很大程度上依赖于世界模型的序列建模和生成能力。然而,为复杂的未知环境构建一个完全准确的模型几乎是不可能的。模型与现实之间的差异可能导致智能体追求虚拟目标,从而在真实环境中表现不佳。将随机噪声引入基于模型的强化学习已被证明是有益的。在这项工作中,我们引入了基于随机Transformer的世界模型(STORM),这是一种高效的世界模型架构,结合了Transformer强大的序列建模和生成能力以及变分自编码器的随机特性。STORM在Atari 100k基准测试中达到了126.7%的平均人类性能,在不使用前瞻搜索技术的最先进方法中创下了新纪录。此外,在单个NVIDIA GeForce RTX 3090显卡上,仅需4.3小时即可训练一个具有1.85小时实时交互经验的智能体,展示了相较于先前方法的更高效率。

英文摘要:

Recently, model-based reinforcement learning algorithms have demonstrated remarkable efficacy in visual input environments. These approaches begin by constructing a parameterized simulation world model of the real environment through self-supervised learning. By leveraging the imagination of the world model, the agent's policy is enhanced without the constraints of sampling from the real environment. The performance of these algorithms heavily relies on the sequence modeling and generation capabilities of the world model. However, constructing a perfectly accurate model of a complex unknown environment is nearly impossible. Discrepancies between the model and reality may cause the agent to pursue virtual goals, resulting in subpar performance in the real environment. Introducing random noise into model-based reinforcement learning has been proven beneficial. In this work, we introduce Stochastic Transformer-based wORld Model (STORM), an efficient world model architecture that combines the strong sequence modeling and generation capabilities of Transformers with the stochastic nature of variational autoencoders. STORM achieves a mean human performance of $126.7\%$ on the Atari $100$k benchmark, setting a new record among state-of-the-art methods that do not employ lookahead search techniques. Moreover, training an agent with $1.85$ hours of real-time interaction experience on a single NVIDIA GeForce RTX 3090 graphics card requires only $4.3$ hours, showcasing improved efficiency compared to previous methodologies.

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。

↑