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 1906.09237cs.LGcs.AIstat.ML
收藏

RL的生成环境模型信念状态的形成

Shaping Belief States with Generative Environment Models for RL

Google DeepMind(谷歌DeepMind)

Karol Gregor, Danilo Jimenez Rezende, Frederic Besse, Yan Wu, Hamza Merzic, Aaron van den Oord

摘要:

当智能体与复杂的环境交互时,它们必须形成并保持关于该环境的相关方面的信念。我们提出了一种方法来有效地训练表达生成模型在复杂的环境。我们表明,具有表达性生成模型的预测算法可以在视觉丰富和动态的3D环境中形成稳定的信念状态。更确切地说,我们表明,学习表示捕捉环境的布局以及代理的位置和方向。我们的实验表明,该模型大大提高了数据效率的一些强化学习(RL)的任务相比,强大的无模型基线代理。我们发现,预测未来的多个步骤(过冲),结合表达生成模型,对于稳定的表示出现至关重要。在实践中,在RL中使用表达生成模型在计算上是昂贵的,我们提出了一个方案来减少这种计算负担,使我们能够构建与无模型基线竞争的代理。

英文摘要:

When agents interact with a complex environment, they must form and maintain beliefs about the relevant aspects of that environment. We propose a way to efficiently train expressive generative models in complex environments. We show that a predictive algorithm with an expressive generative model can form stable belief-states in visually rich and dynamic 3D environments. More precisely, we show that the learned representation captures the layout of the environment as well as the position and orientation of the agent. Our experiments show that the model substantially improves data-efficiency on a number of reinforcement learning (RL) tasks compared with strong model-free baseline agents. We find that predicting multiple steps into the future (overshooting), in combination with an expressive generative model, is critical for stable representations to emerge. In practice, using expressive generative models in RL is computationally expensive and we propose a scheme to reduce this computational burden, allowing us to build agents that are competitive with model-free baselines.

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。

↑