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 2008.12775cs.LGcs.AIcs.ROstat.ML
收藏

基于模型的随机值梯度连续强化学习

On the model-based stochastic value gradient for continuous reinforcement learning

Facebook AI Research(脸书人工智能研究院);NYU(纽约大学)

Brandon Amos, Samuel Stanton, Denis Yarats, Andrew Gordon Wilson

摘要:

十多年来,基于模型的强化学习一直被视为一种利用基于控制的领域知识来提高强化学习代理的样本效率的方法。虽然基于模型的代理在概念上很有吸引力,但它们的策略在最终奖励方面往往落后于无模型代理,特别是在非平凡的环境中。作为回应,研究人员提出了基于模型的代理越来越复杂的组件,从概率动力学模型的集合,以减少模型错误的算法。在这种趋势的逆转,我们表明,简单的基于模型的代理可以从现有的想法,不仅匹配,但优于最先进的无模型代理的样本效率和最终奖励。我们发现,一个无模型的软值估计的政策评估和基于模型的随机值梯度的政策改进是一个有效的组合,实现国家的最先进的结果,在高维人形控制任务,大多数基于模型的代理是无法解决的。我们的研究结果表明,基于模型的政策评估值得密切关注。

英文摘要:

For over a decade, model-based reinforcement learning has been seen as a way to leverage control-based domain knowledge to improve the sample-efficiency of reinforcement learning agents. While model-based agents are conceptually appealing, their policies tend to lag behind those of model-free agents in terms of final reward, especially in non-trivial environments. In response, researchers have proposed model-based agents with increasingly complex components, from ensembles of probabilistic dynamics models, to heuristics for mitigating model error. In a reversal of this trend, we show that simple model-based agents can be derived from existing ideas that not only match, but outperform state-of-the-art model-free agents in terms of both sample-efficiency and final reward. We find that a model-free soft value estimate for policy evaluation and a model-based stochastic value gradient for policy improvement is an effective combination, achieving state-of-the-art results on a high-dimensional humanoid control task, which most model-based agents are unable to solve. Our findings suggest that model-based policy evaluation deserves closer attention.

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。

↑