arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2201.12122cs.LGcs.AIcs.CL
收藏

维基百科能帮助离线强化学习吗?

Can Wikipedia Help Offline Reinforcement Learning?

The University of Tokyo(东京大学);Yale University(耶鲁大学);Google Brain(谷歌大脑)

Machel Reid, Yutaro Yamada, Shixiang Shane Gu

摘要:

由于缺乏大规模的现成数据集以及不同环境之间的可移植性差异很大,因此微调强化学习(RL)模型一直具有挑战性。最近的工作着眼于从序列建模的角度处理离线RL,由于引入了Transformer架构,结果有所改善。然而,当模型从头开始训练时,它的收敛速度很慢。在本文中,我们希望利用这种强化学习的公式作为序列建模,并研究在离线RL任务(控制,游戏)上进行微调时,预训练序列模型在其他领域(视觉,语言)上的可移植性。为此,我们还提出了技术,以提高这些域之间的传输。结果显示,在各种环境下,收敛速度和奖励方面的性能都有一致的提高,将训练速度提高了3- 6倍,并在使用维基百科预训练和GPT 2语言模型的各种任务中实现了最先进的性能。我们希望这项工作不仅能揭示利用通用序列建模技术和RL预训练模型的潜力,还能激发未来在完全不同领域的生成建模任务之间共享知识的工作。

英文摘要:

Fine-tuning reinforcement learning (RL) models has been challenging because of a lack of large scale off-the-shelf datasets as well as high variance in transferability among different environments. Recent work has looked at tackling offline RL from the perspective of sequence modeling with improved results as result of the introduction of the Transformer architecture. However, when the model is trained from scratch, it suffers from slow convergence speeds. In this paper, we look to take advantage of this formulation of reinforcement learning as sequence modeling and investigate the transferability of pre-trained sequence models on other domains (vision, language) when finetuned on offline RL tasks (control, games). To this end, we also propose techniques to improve transfer between these domains. Results show consistent performance gains in terms of both convergence speed and reward on a variety of environments, accelerating training by 3-6x and achieving state-of-the-art performance in a variety of tasks using Wikipedia-pretrained and GPT2 language models. We hope that this work not only brings light to the potentials of leveraging generic sequence modeling techniques and pre-trained models for RL, but also inspires future work on sharing knowledge between generative modeling tasks of completely different domains.

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。

↑