We consider an agent's uncertainty about its environment and the problem of generalizing this uncertainty across states. Specifically, we focus on the problem of exploration in non-tabular reinforcement learning. Drawing inspiration from the intrinsic motivation literature, we use density models to measure uncertainty, and propose a novel algorithm for deriving a pseudo-count from an arbitrary density model. This technique enables us to generalize count-based exploration algorithms to the non-tabular case. We apply our ideas to Atari 2600 games, providing sensible pseudo-counts from raw pixels. We transform these pseudo-counts into exploration bonuses and obtain significantly improved exploration in a number of hard games, including the infamously difficult MONTEZUMA'S REVENGE.

In this paper we argue for the fundamental importance of the value distribution: the distribution of the random return received by a reinforcement learning agent. This is in contrast to the common approach to reinforcement learning which models the expectation of this return, or value. Although there is an established body of literature studying the value distribution, thus far it has always been used for a specific purpose such as implementing risk-aware behaviour. We begin with theoretical results in both the policy evaluation and control settings, exposing a significant distributional instability in the latter. We then use the distributional perspective to design a new algorithm which applies Bellman's equation to the learning of approximate value distributions. We evaluate our algorithm using the suite of games from the Arcade Learning Environment. We obtain both state-of-the-art results and anecdotal evidence demonstrating the importance of the value distribution in approximate reinforcement learning. Finally, we combine theoretical and empirical evidence to highlight the ways in which the value distribution impacts learning in the approximate setting.

我们提出了在表格，依赖阶段的，情节的马尔可夫决策过程中使用贝叶斯-UCBVI算法进行增强学习的：Kaufmann等人的贝叶斯-UCB算法的自然扩展。 （2012年）用于多军匪徒。我们的方法将Q值函数后部的分位数用作最佳Q值函数上的上限。对于贝叶斯-UCBVI，我们证明了一个遗憾的是$ \ wideTilde {o}（\ sqrt {h^3sat}）$，其中$ h $是一集的长度，$ s $是$ s $的数量，$ a $ a $动作数量，$ t $情节数，与$ \ omega（\ sqrt {h^3sat}）$符合poly-$ \ $ \ log $ enter $ h，s，s，a，a，a，a，a ，适用于足够大的$ t $的t $。据我们所知，这是第一种获得对地平线$ h $（和$ s $）的最佳依赖性的算法，而无需涉及伯恩斯坦的奖金或噪音。对于我们的分析而言，至关重要的是一种新的细粒抗浓缩，以具有独立感兴趣的加权dirichlet总和。然后，我们解释了如何轻松地将贝叶斯-UCBVI延伸到表格环境之外，从而在我们的算法和贝叶斯引导之间表现出牢固的联系（Rubin，1981）。

增强学习（RL）研究领域非常活跃，并具有重要的新贡献；特别是考虑到深RL（DRL）的新兴领域。但是，仍然需要解决许多科学和技术挑战，其中我们可以提及抽象行动的能力或在稀疏回报环境中探索环境的难以通过内在动机（IM）来解决的。我们建议通过基于信息理论的新分类法调查这些研究工作：我们在计算上重新审视了惊喜，新颖性和技能学习的概念。这使我们能够确定方法的优势和缺点，并展示当前的研究前景。我们的分析表明，新颖性和惊喜可以帮助建立可转移技能的层次结构，从而进一步抽象环境并使勘探过程更加健壮。

深度强化学习（DRL）和深度多机构的强化学习（MARL）在包括游戏AI，自动驾驶汽车，机器人技术等各种领域取得了巨大的成功。但是，众所周知，DRL和Deep MARL代理的样本效率低下，即使对于相对简单的问题设置，通常也需要数百万个相互作用，从而阻止了在实地场景中的广泛应用和部署。背后的一个瓶颈挑战是众所周知的探索问题，即如何有效地探索环境和收集信息丰富的经验，从而使政策学习受益于最佳研究。在稀疏的奖励，吵闹的干扰，长距离和非平稳的共同学习者的复杂环境中，这个问题变得更加具有挑战性。在本文中，我们对单格和多代理RL的现有勘探方法进行了全面的调查。我们通过确定有效探索的几个关键挑战开始调查。除了上述两个主要分支外，我们还包括其他具有不同思想和技术的著名探索方法。除了算法分析外，我们还对一组常用基准的DRL进行了全面和统一的经验比较。根据我们的算法和实证研究，我们终于总结了DRL和Deep Marl中探索的公开问题，并指出了一些未来的方向。

主动推断是建模生物学和人造药物的行为的概率框架，该框架源于最小化自由能的原理。近年来，该框架已成功地应用于各种情况下，其目标是最大程度地提高奖励，提供可比性，有时甚至是卓越的性能与替代方法。在本文中，我们通过演示如何以及何时进行主动推理代理执行最佳奖励的动作来阐明奖励最大化和主动推断之间的联系。确切地说，我们展示了主动推理为Bellman方程提供最佳解决方案的条件 - 这种公式是基于模型的增强学习和控制的几种方法。在部分观察到的马尔可夫决策过程中，标准的主动推理方案可以为计划视野1的最佳动作产生最佳动作，但不能超越。相比之下，最近开发的递归活跃推理方案（复杂的推理）可以在任何有限的颞范围内产生最佳作用。我们通过讨论主动推理和强化学习之间更广泛的关系来补充分析。

由于策略梯度定理导致的策略设置存在各种理论上 - 声音策略梯度算法，其为梯度提供了简化的形式。然而，由于存在多重目标和缺乏明确的脱助政策政策梯度定理，截止策略设置不太明确。在这项工作中，我们将这些目标统一到一个违规目标，并为此统一目标提供了政策梯度定理。推导涉及强调的权重和利息职能。我们显示多种策略来近似梯度，以识别权重（ACE）称为Actor评论家的算法。我们证明了以前（半梯度）脱离政策演员 - 评论家 - 特别是offpac和DPG - 收敛到错误的解决方案，而Ace找到最佳解决方案。我们还强调为什么这些半梯度方法仍然可以在实践中表现良好，表明ace中的方差策略。我们经验研究了两个经典控制环境的若干ACE变体和基于图像的环境，旨在说明每个梯度近似的权衡。我们发现，通过直接逼近强调权重，ACE在所有测试的所有设置中执行或优于offpac。

Adequately assigning credit to actions for future outcomes based on their contributions is a long-standing open challenge in Reinforcement Learning. The assumptions of the most commonly used credit assignment method are disadvantageous in tasks where the effects of decisions are not immediately evident. Furthermore, this method can only evaluate actions that have been selected by the agent, making it highly inefficient. Still, no alternative methods have been widely adopted in the field. Hindsight Credit Assignment is a promising, but still unexplored candidate, which aims to solve the problems of both long-term and counterfactual credit assignment. In this thesis, we empirically investigate Hindsight Credit Assignment to identify its main benefits, and key points to improve. Then, we apply it to factored state representations, and in particular to state representations based on the causal structure of the environment. In this setting, we propose a variant of Hindsight Credit Assignment that effectively exploits a given causal structure. We show that our modification greatly decreases the workload of Hindsight Credit Assignment, making it more efficient and enabling it to outperform the baseline credit assignment method on various tasks. This opens the way to other methods based on given or learned causal structures.

我们介绍了一种改进政策改进的方法，该方法在基于价值的强化学习（RL）的贪婪方法与基于模型的RL的典型计划方法之间进行了插值。新方法建立在几何视野模型（GHM，也称为伽马模型）的概念上，该模型对给定策略的折现状态验证分布进行了建模。我们表明，我们可以通过仔细的基本策略GHM的仔细组成，而无需任何其他学习，可以评估任何非马尔科夫策略，以固定的概率在一组基本马尔可夫策略之间切换。然后，我们可以将广义政策改进（GPI）应用于此类非马尔科夫政策的收集，以获得新的马尔可夫政策，通常将其表现优于其先驱。我们对这种方法提供了彻底的理论分析，开发了转移和标准RL的应用，并在经验上证明了其对标准GPI的有效性，对充满挑战的深度RL连续控制任务。我们还提供了GHM培训方法的分析，证明了关于先前提出的方法的新型收敛结果，并显示了如何在深度RL设置中稳定训练这些模型。

Deep reinforcement learning is poised to revolutionise the field of AI and represents a step towards building autonomous systems with a higher level understanding of the visual world. Currently, deep learning is enabling reinforcement learning to scale to problems that were previously intractable, such as learning to play video games directly from pixels. Deep reinforcement learning algorithms are also applied to robotics, allowing control policies for robots to be learned directly from camera inputs in the real world. In this survey, we begin with an introduction to the general field of reinforcement learning, then progress to the main streams of value-based and policybased methods. Our survey will cover central algorithms in deep reinforcement learning, including the deep Q-network, trust region policy optimisation, and asynchronous advantage actor-critic. In parallel, we highlight the unique advantages of deep neural networks, focusing on visual understanding via reinforcement learning. To conclude, we describe several current areas of research within the field.

We revisit the domain of off-policy policy optimization in RL from the perspective of coordinate ascent. One commonly-used approach is to leverage the off-policy policy gradient to optimize a surrogate objective -- the total discounted in expectation return of the target policy with respect to the state distribution of the behavior policy. However, this approach has been shown to suffer from the distribution mismatch issue, and therefore significant efforts are needed for correcting this mismatch either via state distribution correction or a counterfactual method. In this paper, we rethink off-policy learning via Coordinate Ascent Policy Optimization (CAPO), an off-policy actor-critic algorithm that decouples policy improvement from the state distribution of the behavior policy without using the policy gradient. This design obviates the need for distribution correction or importance sampling in the policy improvement step of off-policy policy gradient. We establish the global convergence of CAPO with general coordinate selection and then further quantify the convergence rates of several instances of CAPO with popular coordinate selection rules, including the cyclic and the randomized variants of CAPO. We then extend CAPO to neural policies for a more practical implementation. Through experiments, we demonstrate that CAPO provides a competitive approach to RL in practice.

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.

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.

Atari games have been a long-standing benchmark in the reinforcement learning (RL) community for the past decade. This benchmark was proposed to test general competency of RL algorithms. Previous work has achieved good average performance by doing outstandingly well on many games of the set, but very poorly in several of the most challenging games. We propose Agent57, the first deep RL agent that outperforms the standard human benchmark on all 57 Atari games. To achieve this result, we train a neural network which parameterizes a family of policies ranging from very exploratory to purely exploitative. We propose an adaptive mechanism to choose which policy to prioritize throughout the training process. Additionally, we utilize a novel parameterization of the architecture that allows for more consistent and stable learning.

自成立以来，建立在广泛任务中表现出色的普通代理的任务一直是强化学习的重要目标。这个问题一直是对Alarge工作体系的研究的主题，并且经常通过观察Atari 57基准中包含的广泛范围环境的分数来衡量的性能。 Agent57是所有57场比赛中第一个超过人类基准的代理商，但这是以数据效率差的代价，需要实现近800亿帧的经验。以Agent57为起点，我们采用了各种各样的形式，以降低超过人类基线所需的经验200倍。在减少数据制度和Propose有效的解决方案时，我们遇到了一系列不稳定性和瓶颈，以构建更强大，更有效的代理。我们还使用诸如Muesli和Muzero之类的高性能方法证明了竞争性的性能。 TOOUR方法的四个关键组成部分是（1）近似信任区域方法，该方法可以从TheOnline网络中稳定引导，（2）损失和优先级的归一化方案，在学习具有广泛量表的一组值函数时，可以提高鲁棒性， （3）改进的体系结构采用了NFNET的技术技术来利用更深的网络而无需标准化层，并且（4）政策蒸馏方法可使瞬时贪婪的策略加班。

In this paper, we consider the problem of adjusting the exploration rate when using value-of-information-based exploration. We do this by converting the value-of-information optimization into a problem of finding equilibria of a flow for a changing exploration rate. We then develop an efficient path-following scheme for converging to these equilibria and hence uncovering optimal action-selection policies. Under this scheme, the exploration rate is automatically adapted according to the agent's experiences. Global convergence is theoretically assured. We first evaluate our exploration-rate adaptation on the Nintendo GameBoy games Centipede and Millipede. We demonstrate aspects of the search process. We show that our approach yields better policies in fewer episodes than conventional search strategies relying on heuristic, annealing-based exploration-rate adjustments. We then illustrate that these trends hold for deep, value-of-information-based agents that learn to play ten simple games and over forty more complicated games for the Nintendo GameBoy system. Performance either near or well above the level of human play is observed.

我们介绍了一种普遍的策略，可实现有效的多目标勘探。它依赖于adagoal，一种基于简单约束优化问题的新的目标选择方案，其自适应地针对目标状态，这既不是太困难也不是根据代理目前的知识达到的。我们展示了Adagoal如何用于解决学习$ \ epsilon $ -optimal的目标条件的政策，以便在$ L $ S_0 $ S_0 $奖励中获得的每一个目标状态，以便在$ S_0 $中获取。免费马尔可夫决策过程。在标准的表格外壳中，我们的算法需要$ \ tilde {o}（l ^ 3 s a \ epsilon ^ { - 2}）$探索步骤，这几乎很少最佳。我们还容易在线性混合Markov决策过程中实例化Adagoal，其产生具有线性函数近似的第一目标导向的PAC保证。除了强大的理论保证之外，迈克纳队以现有方法的高级别算法结构为锚定，为目标条件的深度加固学习。

由政策引起的马尔可夫链的混合时间限制了现实世界持续学习场景中的性能。然而，混合时间对持续增强学习学习（RL）的影响仍然是曝光率。在本文中，我们表征了长期兴趣的问题，以通过混合时间调用可扩展的MDP来发展持续的RL。特别是，我们建立可扩展的MDP具有与问题的大小相等的混合时间。我们继续证明，多项式混合时间对现有方法产生显着困难，并提出了一种基于模型的算法，通过新颖的引导程序直接优化平均奖励来加速学习。最后，我们对我们提出的方法进行了实证遗憾分析，展示了对基线的清晰改进，以及如何使用可缩放的MDP来分析RL算法作为混合时间规模。

深度Q网络（DQN）通过将深度学习（DL）与加强学习（RL）组合，这已经注意到，已经注意到所获取的数据的分布在训练过程中将改变。 DQN发现此属性可能会导致培训不稳定，因此它提出了处理财产缺点的有效方法。而不是专注于不利的方面，我们发现RL为缓解估计的数据分布与地面真理数据分布之间的差距，而是在监督学习（SL）未能这样做的情况下是至关重要的。从这种新的角度来看，我们将称为广义策略迭代（GPI）的RL的基本范例扩展到更广泛的版本中，该版本称为广义数据分发迭代（GDI）。我们看到大规模的RL算法和技术可以统一到GDI范式，这可以被认为是GDI的特殊情况之一。我们提供理论证明，为什么GDI比GPI更好，以及它如何运作。已经提出了基于GDI的几种实用算法来验证它的有效性和广泛性。经验实验证明我们在街机学习环境（ALE）上的最先进的（SOTA）性能，其中我们的算法达到了9620.98％的平均值人类规范化得分（HNS），1146.39％的中位数HNS和22人的世界历史突破（HWRB ）仅使用仅200M的训练框架。我们的工作旨在引领RL研究进入征服人类世界纪录的旅程，并在表现和效率上寻求真正的超人代理。

由于数据量增加，金融业的快速变化已经彻底改变了数据处理和数据分析的技术，并带来了新的理论和计算挑战。与古典随机控制理论和解决财务决策问题的其他分析方法相比，解决模型假设的财务决策问题，强化学习（RL）的新发展能够充分利用具有更少模型假设的大量财务数据并改善复杂的金融环境中的决策。该调查纸目的旨在审查最近的资金途径的发展和使用RL方法。我们介绍了马尔可夫决策过程，这是许多常用的RL方法的设置。然后引入各种算法，重点介绍不需要任何模型假设的基于价值和基于策略的方法。连接是用神经网络进行的，以扩展框架以包含深的RL算法。我们的调查通过讨论了这些RL算法在金融中各种决策问题中的应用，包括最佳执行，投资组合优化，期权定价和对冲，市场制作，智能订单路由和Robo-Awaring。