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 2512.19366cs.AIcs.LG
收藏

通过策略梯度方法学习通用策略

Learning General Policies with Policy Gradient Methods

Linköping University(利厄普大学);Universitat Pompeu Fabra(庞培法拉大学);RWTH Aachen University(亚琛工业大学)

Simon Ståhlberg, Blai Bonet, Hector Geffner

AI总结:

本文通过策略梯度方法结合图神经网络,实现了与组合方法相当的泛化能力,同时避免了可扩展性瓶颈和特征池的使用。

摘要:

尽管强化学习方法在许多设定中取得了显著成果,但泛化能力,即能够产生可靠且系统化的方式产生策略的能力,仍然是一个挑战。在经典规划中,已经正式解决了泛化问题,通过组合方法学习了能够泛化所有给定领域实例的可证明正确的策略。本研究的目标是将这两个研究领域结合起来,以揭示在何种条件下(深度)强化学习方法,特别是策略优化方法,可以用来学习像组合方法一样泛化的策略。我们借鉴了之前组合和深度学习方法中的经验,并以方便的方式进行了扩展。从前者来看,我们将策略建模为状态转移分类器,因为(地面)动作不是通用的,且在实例之间变化。从后者来看,我们使用图神经网络(GNNs)来处理关系结构,以表示规划状态上的价值函数,在我们的情况下是策略。有了这些成分,我们发现可以使用actor-critic方法学习几乎与组合方法获得的策略一样泛化的策略,同时避免了可扩展性瓶颈和特征池的使用。此外,所考虑基准测试中深度强化学习方法的局限性与深度学习或强化学习算法关系不大,而是源于对GNNs的表达限制以及最优性和泛化之间的权衡(在某些领域中通用策略不能最优)。这两个限制通过在不改变基本DRL方法的情况下添加派生谓词和替代成本结构来解决。

英文摘要:

While reinforcement learning methods have delivered remarkable results in a number of settings, generalization, i.e., the ability to produce policies that generalize in a reliable and systematic way, has remained a challenge. The problem of generalization has been addressed formally in classical planning where provable correct policies that generalize over all instances of a given domain have been learned using combinatorial methods. The aim of this work is to bring these two research threads together to illuminate the conditions under which (deep) reinforcement learning approaches, and in particular, policy optimization methods, can be used to learn policies that generalize like combinatorial methods do. We draw on lessons learned from previous combinatorial and deep learning approaches, and extend them in a convenient way. From the former, we model policies as state transition classifiers, as (ground) actions are not general and change from instance to instance. From the latter, we use graph neural networks (GNNs) adapted to deal with relational structures for representing value functions over planning states, and in our case, policies. With these ingredients in place, we find that actor-critic methods can be used to learn policies that generalize almost as well as those obtained using combinatorial approaches while avoiding the scalability bottleneck and the use of feature pools. Moreover, the limitations of the DRL methods on the benchmarks considered have little to do with deep learning or reinforcement learning algorithms, and result from the well-understood expressive limitations of GNNs, and the tradeoff between optimality and generalization (general policies cannot be optimal in some domains). Both of these limitations are addressed without changing the basic DRL methods by adding derived predicates and an alternative cost structure to optimize.

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。

↑