我们介绍了Smianile过滤仿制学习（QFIL），这是一种用于离线强化学习的新型政策改进操作员。QFIL通过在脱机数据集的过滤版本上运行模仿学习来执行策略改进。过滤过程删除了$ s，其估计的q值低于给定分位于通过从行为策略采样动作引起的值的推送分布。推轴Q分布和产生的值函数分位数的定义是我们方法的主要贡献。我们证明QFIL为我们提供了一种安全的政策改进步骤，函数近似，分位式的选择提供了自然的超参数，以折衷偏差和改进步骤的差异。凭经验，我们执行一个合成实验，说明QFIL如何有效地进行偏差方差权衡，并且我们看到QFil在D4RL基准上表现良好。

大多数前往离线强化学习（RL）的方法都采取了一种迭代演员 - 批评批评，涉及违规评估。在本文中，我们展示了使用行为政策的政策Q估计来令人惊讶地执行一步的Q估计，从而简单地执行一个受限制/正规化的政策改进的步骤。该一步算法在大部分D4RL基准测试中击败了先前报告的迭代算法的结果。一步基线实现了这种强劲的性能，同时对超公数更简单，更强大而不是先前提出的迭代算法。我们认为迭代方法的表现相对较差是在违反政策评估中固有的高方差，并通过对这些估计的重复优化的政策进行放大。此外，我们假设一步算法的强大性能是由于环境和行为政策中有利结构的组合。

离线增强学习（RL）将经典RL算法的范式扩展到纯粹从静态数据集中学习，而无需在学习过程中与基础环境进行交互。离线RL的一个关键挑战是政策培训的不稳定，这是由于离线数据的分布与学习政策的未结束的固定状态分配之间的不匹配引起的。为了避免分配不匹配的有害影响，我们将当前政策的未静置固定分配正规化在政策优化过程中的离线数据。此外，我们训练动力学模型既实施此正规化，又可以更好地估计当前策略的固定分布，从而减少了分布不匹配引起的错误。在各种连续控制的离线RL数据集中，我们的方法表示竞争性能，从而验证了我们的算法。该代码公开可用。

我们根据相对悲观主义的概念，在数据覆盖不足的情况下提出了经过对抗训练的演员评论家（ATAC），这是一种新的无模型算法（RL）。 ATAC被设计为两人Stackelberg游戏：政策演员与受对抗训练的价值评论家竞争，后者发现参与者不如数据收集行为策略的数据一致方案。我们证明，当演员在两人游戏中不后悔时，运行ATAC会产生一项政策，证明1）在控制悲观程度的各种超级参数上都超过了行为政策，而2）与最佳竞争。 policy covered by data with appropriately chosen hyperparameters.与现有作品相比，尤其是我们的框架提供了一般函数近似的理论保证，也提供了可扩展到复杂环境和大型数据集的深度RL实现。在D4RL基准测试中，ATAC在一系列连续的控制任务上始终优于最先进的离线RL算法。

Effectively leveraging large, previously collected datasets in reinforcement learning (RL) is a key challenge for large-scale real-world applications. Offline RL algorithms promise to learn effective policies from previously-collected, static datasets without further interaction. However, in practice, offline RL presents a major challenge, and standard off-policy RL methods can fail due to overestimation of values induced by the distributional shift between the dataset and the learned policy, especially when training on complex and multi-modal data distributions. In this paper, we propose conservative Q-learning (CQL), which aims to address these limitations by learning a conservative Q-function such that the expected value of a policy under this Q-function lower-bounds its true value. We theoretically show that CQL produces a lower bound on the value of the current policy and that it can be incorporated into a policy learning procedure with theoretical improvement guarantees. In practice, CQL augments the standard Bellman error objective with a simple Q-value regularizer which is straightforward to implement on top of existing deep Q-learning and actor-critic implementations. On both discrete and continuous control domains, we show that CQL substantially outperforms existing offline RL methods, often learning policies that attain 2-5 times higher final return, especially when learning from complex and multi-modal data distributions.Preprint. Under review.

Behavior constrained policy optimization has been demonstrated to be a successful paradigm for tackling Offline Reinforcement Learning. By exploiting historical transitions, a policy is trained to maximize a learned value function while constrained by the behavior policy to avoid a significant distributional shift. In this paper, we propose our closed-form policy improvement operators. We make a novel observation that the behavior constraint naturally motivates the use of first-order Taylor approximation, leading to a linear approximation of the policy objective. Additionally, as practical datasets are usually collected by heterogeneous policies, we model the behavior policies as a Gaussian Mixture and overcome the induced optimization difficulties by leveraging the LogSumExp's lower bound and Jensen's Inequality, giving rise to a closed-form policy improvement operator. We instantiate offline RL algorithms with our novel policy improvement operators and empirically demonstrate their effectiveness over state-of-the-art algorithms on the standard D4RL benchmark.

深度强化学习（RL）导致了许多最近和开创性的进步。但是，这些进步通常以培训的基础体系结构的规模增加以及用于训练它们的RL算法的复杂性提高，而均以增加规模的成本。这些增长反过来又使研究人员更难迅速原型新想法或复制已发表的RL算法。为了解决这些问题，这项工作描述了ACME，这是一个用于构建新型RL算法的框架，这些框架是专门设计的，用于启用使用简单的模块化组件构建的代理，这些组件可以在各种执行范围内使用。尽管ACME的主要目标是为算法开发提供一个框架，但第二个目标是提供重要或最先进算法的简单参考实现。这些实现既是对我们的设计决策的验证，也是对RL研究中可重复性的重要贡献。在这项工作中，我们描述了ACME内部做出的主要设计决策，并提供了有关如何使用其组件来实施各种算法的进一步详细信息。我们的实验为许多常见和最先进的算法提供了基准，并显示了如何为更大且更复杂的环境扩展这些算法。这突出了ACME的主要优点之一，即它可用于实现大型，分布式的RL算法，这些算法可以以较大的尺度运行，同时仍保持该实现的固有可读性。这项工作提出了第二篇文章的版本，恰好与模块化的增加相吻合，对离线，模仿和从演示算法学习以及作为ACME的一部分实现的各种新代理。

依赖于太多的实验来学习良好的行动，目前的强化学习（RL）算法在现实世界的环境中具有有限的适用性，这可能太昂贵，无法探索探索。我们提出了一种批量RL算法，其中仅使用固定的脱机数据集来学习有效策略，而不是与环境的在线交互。批量RL中的有限数据产生了在培训数据中不充分表示的状态/行动的价值估计中的固有不确定性。当我们的候选政策从生成数据的候选政策发散时，这导致特别严重的外推。我们建议通过两个直接的惩罚来减轻这个问题：减少这种分歧的政策限制和减少过于乐观估计的价值约束。在全面的32个连续动作批量RL基准测试中，我们的方法对最先进的方法进行了比较，无论如何收集离线数据如何。

In offline reinforcement learning (RL), a learner leverages prior logged data to learn a good policy without interacting with the environment. A major challenge in applying such methods in practice is the lack of both theoretically principled and practical tools for model selection and evaluation. To address this, we study the problem of model selection in offline RL with value function approximation. The learner is given a nested sequence of model classes to minimize squared Bellman error and must select among these to achieve a balance between approximation and estimation error of the classes. We propose the first model selection algorithm for offline RL that achieves minimax rate-optimal oracle inequalities up to logarithmic factors. The algorithm, ModBE, takes as input a collection of candidate model classes and a generic base offline RL algorithm. By successively eliminating model classes using a novel one-sided generalization test, ModBE returns a policy with regret scaling with the complexity of the minimally complete model class. In addition to its theoretical guarantees, it is conceptually simple and computationally efficient, amounting to solving a series of square loss regression problems and then comparing relative square loss between classes. We conclude with several numerical simulations showing it is capable of reliably selecting a good model class.

In reinforcement learning an agent interacts with the environment by taking actions and observing the next state and reward. When sampled probabilistically, these state transitions, rewards, and actions can all induce randomness in the observed long-term return. Traditionally, reinforcement learning algorithms average over this randomness to estimate the value function. In this paper, we build on recent work advocating a distributional approach to reinforcement learning in which the distribution over returns is modeled explicitly instead of only estimating the mean. That is, we examine methods of learning the value distribution instead of the value function. We give results that close a number of gaps between the theoretical and algorithmic results given by Bellemare, . First, we extend existing results to the approximate distribution setting. Second, we present a novel distributional reinforcement learning algorithm consistent with our theoretical formulation. Finally, we evaluate this new algorithm on the Atari 2600 games, observing that it significantly outperforms many of the recent improvements on DQN, including the related distributional algorithm C51.

我们研究了离线加强学习（RL）的代表性学习，重点是离线政策评估（OPE）的重要任务。最近的工作表明，与监督的学习相反，Q功能的可实现性不足以学习。样品效率OPE的两个足够条件是Bellman的完整性和覆盖范围。先前的工作通常假设给出满足这些条件的表示形式，结果大多是理论上的。在这项工作中，我们提出了BCRL，该BCRL直接从数据中吸取了近似线性的贝尔曼完整表示，并具有良好的覆盖范围。通过这种学识渊博的表示，我们使用最小平方策略评估（LSPE）执行OPE，并在我们学习的表示中具有线性函数。我们提出了端到端的理论分析，表明我们的两阶段算法享有多项式样本复杂性，该算法在所考虑的丰富类别中提供了一些表示形式，这是线性的贝尔曼完成。从经验上讲，我们广泛评估了我们的DeepMind Control Suite的具有挑战性的基于图像的连续控制任务。我们显示我们的表示能够与针对非政策RL开发的先前表示的学习方法（例如Curl，SPR）相比，可以更好地使用OPE。 BCRL使用最先进的方法拟合Q评估（FQE）实现竞争性OPE误差，并在评估超出初始状态分布的评估时击败FQE。我们的消融表明，我们方法的线性铃铛完整和覆盖范围都至关重要。

Off-policy reinforcement learning aims to leverage experience collected from prior policies for sample-efficient learning. However, in practice, commonly used off-policy approximate dynamic programming methods based on Q-learning and actor-critic methods are highly sensitive to the data distribution, and can make only limited progress without collecting additional on-policy data. As a step towards more robust off-policy algorithms, we study the setting where the off-policy experience is fixed and there is no further interaction with the environment. We identify bootstrapping error as a key source of instability in current methods. Bootstrapping error is due to bootstrapping from actions that lie outside of the training data distribution, and it accumulates via the Bellman backup operator. We theoretically analyze bootstrapping error, and demonstrate how carefully constraining action selection in the backup can mitigate it. Based on our analysis, we propose a practical algorithm, bootstrapping error accumulation reduction (BEAR). We demonstrate that BEAR is able to learn robustly from different off-policy distributions, including random and suboptimal demonstrations, on a range of continuous control tasks.

现有的离线增强学习（RL）方法面临一些主要挑战，尤其是学识渊博的政策与行为政策之间的分配转变。离线Meta-RL正在成为应对这些挑战的一种有前途的方法，旨在从一系列任务中学习信息丰富的元基础。然而，如我们的实证研究所示，离线元RL在具有良好数据集质量的任务上的单个任务RL方法可能胜过，这表明必须在“探索”不合时宜的情况下进行精细的平衡。通过遵循元元素和“利用”离线数据集的分配状态行为，保持靠近行为策略。通过这种经验分析的激励，我们探索了基于模型的离线元RL，并具有正则政策优化（MERPO），该策略优化（MERPO）学习了一种用于有效任务结构推理的元模型，并提供了提供信息的元元素，以安全地探索过分分布状态 - 行为。特别是，我们使用保守的政策评估和正规政策改进，设计了一种新的基于元指数的基于元指数的基于元模型的参与者批判性（RAC），作为MERPO的关键构建块作为Merpo的关键构建块；而其中的内在权衡是通过在两个正规机构之间达到正确的平衡来实现的，一个是基于行为政策，另一个基于元政策。从理论上讲，我们学识渊博的政策可以保证对行为政策和元政策都有保证的改进，从而确保通过离线元RL对新任务的绩效提高。实验证实了Merpo优于现有的离线META-RL方法的出色性能。

在训练数据的分布中评估时，学到的模型和政策可以有效地概括，但可以在分布输入输入的情况下产生不可预测且错误的输出。为了避免在部署基于学习的控制算法时分配变化，我们寻求一种机制将代理商限制为类似于受过训练的国家和行动的机制。在控制理论中，Lyapunov稳定性和控制不变的集合使我们能够保证稳定系统周围系统的控制器，而在机器学习中，密度模型使我们能够估算培训数据分布。我们可以将这两个概念结合起来，产生基于学习的控制算法，这些算法仅使用分配动作将系统限制为分布状态？在这项工作中，我们建议通过结合Lyapunov稳定性和密度估计的概念来做到这一点，引入Lyapunov密度模型：控制Lyapunov函数和密度模型的概括，这些函数和密度模型可以保证代理商在其整个轨迹上保持分布的能力。

脱机强化学习 - 从一批数据中学习策略 - 是难以努力的：如果没有制造强烈的假设，它很容易构建实体算法失败的校长。在这项工作中，我们考虑了某些现实世界问题的财产，其中离线强化学习应该有效：行动仅对一部分产生有限的行动。我们正规化并介绍此动作影响规律（AIR）财产。我们进一步提出了一种算法，该算法假定和利用AIR属性，并在MDP满足空气时绑定输出策略的子优相。最后，我们展示了我们的算法在定期保留的两个模拟环境中跨越不同的数据收集策略占据了现有的离线强度学习算法。

离线强化学习（RL）任务要求代理从预先收集的数据集中学习，没有与环境进行进一步的交互。尽管有可能超越行为政策，但基于RL的方法通常是不切实际的，因为培训不稳定并引导外推错误，这始终需要通过在线评估进行仔细的超参数调整。相比之下，离线模仿学习（IL）没有这样的问题，因为它直接在不估计值函数的情况下直接了解策略。然而，IL通常限制在行为政策的能力，并且倾向于从政策混合收集的数据集中学习平庸行为。在本文中，我们的目标是利用IL但缓解这种缺点。观察行为克隆能够使用较少的数据模仿邻近的策略，我们提出\ Textit {课程脱机仿制学习（线圈）}，它利用具有更高回报的自适应邻近策略的体验挑选策略，并提高了当前策略沿课程阶段。在连续控制基准测试中，我们将线圈与基于仿制的和基于RL的方法进行比较，表明它不仅避免了在混合数据集上学习平庸行为，而且甚至与最先进的离线RL方法竞争。

Offline reinforcement learning (RL) promises the ability to learn effective policies solely using existing, static datasets, without any costly online interaction. To do so, offline RL methods must handle distributional shift between the dataset and the learned policy. The most common approach is to learn conservative, or lower-bound, value functions, which underestimate the return of out-of-distribution (OOD) actions. However, such methods exhibit one notable drawback: policies optimized on such value functions can only behave according to a fixed, possibly suboptimal, degree of conservatism. However, this can be alleviated if we instead are able to learn policies for varying degrees of conservatism at training time and devise a method to dynamically choose one of them during evaluation. To do so, in this work, we propose learning value functions that additionally condition on the degree of conservatism, which we dub confidence-conditioned value functions. We derive a new form of a Bellman backup that simultaneously learns Q-values for any degree of confidence with high probability. By conditioning on confidence, our value functions enable adaptive strategies during online evaluation by controlling for confidence level using the history of observations thus far. This approach can be implemented in practice by conditioning the Q-function from existing conservative algorithms on the confidence. We theoretically show that our learned value functions produce conservative estimates of the true value at any desired confidence. Finally, we empirically show that our algorithm outperforms existing conservative offline RL algorithms on multiple discrete control domains.

先前的工作提出了一种简单的加固学习策略（RL）：标签经验，具有该经验中取得的成果，然后模仿重新标记的经验。这些结局条件的模仿学习方法由于其简单性，强大的表现和与监督学习的紧密联系而具有吸引力。但是，尚不清楚这些方法与标准RL目标，最大化如何相关。在本文中，我们证明现有的结果模仿学习方法不一定改善政策。相反，在某些情况下，它们可以减少预期的奖励。但是，我们表明，在某些假设下，简单的修改会导致一种可以保证政策改进的方法。我们的目的不是开发一种全新的方法，而是解释如何使用一种结局条件的模仿学习来最大化奖励。

我们提供了一种通过从域知识或离线数据构建的启发式提供加强学习（RL）算法的框架。 Tabula RAS RL算法需要与顺序决策任务的地平线相比的环境相互作用或计算。使用我们的框架，我们展示了启发式引导的RL如何引导更短的地平次数，可从而解决原始任务。我们的框架可以被视为基于地平线的正则化，用于在有限互动预算下控制RL中的偏差和方差。在理论方面，我们表征了良好启发式的特性及其对RL加速的影响。特别是，我们介绍了一种新颖的启发式的概念，一种启发式，允许RL代理外推超出其先前知识。在实证方面，我们实例化了我们的框架，以加速模拟机器人控制任务和程序生成的游戏中的若干最先进的算法。我们的框架在热启动RL与专家演示或探索数据集中的丰富文学补充，并引入了一种用于将先验知识注入RL的原则方法。

Offline reinforcement learning (RL) refers to the problem of learning policies entirely from a large batch of previously collected data. This problem setting offers the promise of utilizing such datasets to acquire policies without any costly or dangerous active exploration. However, it is also challenging, due to the distributional shift between the offline training data and those states visited by the learned policy. Despite significant recent progress, the most successful prior methods are model-free and constrain the policy to the support of data, precluding generalization to unseen states. In this paper, we first observe that an existing model-based RL algorithm already produces significant gains in the offline setting compared to model-free approaches. However, standard model-based RL methods, designed for the online setting, do not provide an explicit mechanism to avoid the offline setting's distributional shift issue. Instead, we propose to modify the existing model-based RL methods by applying them with rewards artificially penalized by the uncertainty of the dynamics. We theoretically show that the algorithm maximizes a lower bound of the policy's return under the true MDP. We also characterize the trade-off between the gain and risk of leaving the support of the batch data. Our algorithm, Model-based Offline Policy Optimization (MOPO), outperforms standard model-based RL algorithms and prior state-of-the-art model-free offline RL algorithms on existing offline RL benchmarks and two challenging continuous control tasks that require generalizing from data collected for a different task. * equal contribution. † equal advising. Orders randomized.34th Conference on Neural Information Processing Systems (NeurIPS 2020),