Modern reinforcement learning (RL) commonly engages practical problems with large state spaces, where function approximation must be deployed to approximate either the value function or the policy. While recent progresses in RL theory address a rich set of RL problems with general function approximation, such successes are mostly restricted to the single-agent setting. It remains elusive how to extend these results to multi-agent RL, especially due to the new challenges arising from its game-theoretical nature. This paper considers two-player zero-sum Markov Games (MGs). We propose a new algorithm that can provably find the Nash equilibrium policy using a polynomial number of samples, for any MG with low multi-agent Bellman-Eluder dimension -- a new complexity measure adapted from its single-agent version (Jin et al., 2021). A key component of our new algorithm is the exploiter, which facilitates the learning of the main player by deliberately exploiting her weakness. Our theoretical framework is generic, which applies to a wide range of models including but not limited to tabular MGs, MGs with linear or kernel function approximation, and MGs with rich observations.
Efficient reinforcement learning (RL) involves a trade-off between "exploitative" actions that maximise expected reward and "explorative'" ones that sample unvisited states. To encourage exploration, recent approaches proposed adding stochasticity to actions, separating exploration and exploitation phases, or equating reduction in uncertainty with reward. However, these techniques do not necessarily offer entirely systematic approaches making this trade-off. Here we introduce SElective Reinforcement Exploration Network (SEREN) that poses the exploration-exploitation trade-off as a game between…
Nanyang Technological University（南洋理工大学）；The Hong Kong Polytechnic University（香港理工大学）；CAIR, HKISI, CAS（中国科学院香港中文大学信息工程研究所（注：CAIR为Computational Intelligence and Robotics，HKISI为Hong Kong Institute of Science Innovation，CAS为Chinese Academy of Sciences））；Singapore Management University（新加坡管理大学）；University of Nebraska（内布拉斯加大学）
Multi-agent reinforcement learning (MARL) achieves significant empirical successes. However, MARL suffers from the curse of many agents. In this paper, we exploit the symmetry of agents in MARL. In the most generic form, we study a mean-field MARL problem. Such a mean-field MARL is defined on mean-field states, which are distributions that are supported on continuous space. Based on the mean embedding of the distributions, we propose MF-FQI algorithm that solves the mean-field MARL and establishes a non-asymptotic analysis for MF-FQI algorithm. We highlight that MF-FQI algorithm enjoys a "bles…
Self-play, where the algorithm learns by playing against itself without requiring any direct supervision, has become the new weapon in modern Reinforcement Learning (RL) for achieving superhuman performance in practice. However, the majority of exisiting theory in reinforcement learning only applies to the setting where the agent plays against a fixed environment; it remains largely open whether self-play algorithms can be provably effective, especially when it is necessary to manage the exploration/exploitation tradeoff. We study self-play in competitive reinforcement learning under the setti…
In privacy-preserving multi-agent planning, a group of agents attempt to cooperatively solve a multi-agent planning problem while maintaining private their data and actions. Although much work was carried out in this area in past years, its theoretical foundations have not been fully worked out. Specifically, although algorithms with precise privacy guarantees exist, even their most efficient implementations are not fast enough on realistic instances, whereas for practical algorithms no meaningful privacy guarantees exist. Secure-MAFS, a variant of the multi-agent forward search algorithm (MAF…
We propose a provable defense mechanism against backdoor policies in reinforcement learning under subspace trigger assumption. A backdoor policy is a security threat where an adversary publishes a seemingly well-behaved policy which in fact allows hidden triggers. During deployment, the adversary can modify observed states in a particular way to trigger unexpected actions and harm the agent. We assume the agent does not have the resources to re-train a good policy. Instead, our defense mechanism sanitizes the backdoor policy by projecting observed states to a 'safe subspace', estimated from a …
Safety is a crucial necessity in many applications of reinforcement learning (RL), whether robotic, automotive, or medical. Many existing approaches to safe RL rely on receiving numeric safety feedback, but in many cases this feedback can only take binary values; that is, whether an action in a given state is safe or unsafe. This is particularly true when feedback comes from human experts. We therefore consider the problem of provable safe RL when given access to an offline oracle providing binary feedback on the safety of state, action pairs. We provide a novel meta algorithm, SABRE, which ca…
Reinforcement learning (RL) has shown promise as a tool for engineering safe, ethical, or legal behaviour in autonomous agents. Its use typically relies on assigning punishments to state-action pairs that constitute unsafe or unethical choices. Despite this assignment being a crucial step in this approach, however, there has been limited discussion on generalizing the process of selecting punishments and deciding where to apply them. In this paper, we adopt an approach that leverages an existing framework -- the normative supervisor of (Neufeld et al., 2021) -- during training. This normative …
Tsinghua University（清华大学）；University of Washington（华盛顿大学）
While multitask representation learning has become a popular approach in reinforcement learning (RL) to boost the sample efficiency, the theoretical understanding of why and how it works is still limited. Most previous analytical works could only assume that the representation function is already known to the agent or from linear function class, since analyzing general function class representation encounters non-trivial technical obstacles such as generalization guarantee, formulation of confidence bound in abstract function space, etc. However, linear-case analysis heavily relies on the part…
Google Research（谷歌研究院）；Carnegie Mellon University（卡内基梅隆大学）；Cornell University（康奈尔大学）；Princeton University（普林斯顿大学）
We study the problem of representational transfer in RL, where an agent first pretrains in a number of source tasks to discover a shared representation, which is subsequently used to learn a good policy in a \emph{target task}. We propose a new notion of task relatedness between source and target tasks, and develop a novel approach for representational transfer under this assumption. Concretely, we show that given generative access to source tasks, we can discover a representation, using which subsequent linear RL techniques quickly converge to a near-optimal policy in the target task. The sam…
On the Power of Pre-training for Generalization in RL: Provable Benefits and Hardness
Yuanpei College, Peking University（北京大学元培学院）；School of Intelligence Science and Technology, Peking University（北京大学智能学院）；Center for Data Science, Peking University（北京大学数据科学中心）；University of Washington（华盛顿大学）
Microsoft Research（微软研究院）；Princeton University（普林斯顿大学）
Real-world sequential decision making problems commonly involve partial observability, which requires the agent to maintain a memory of history in order to infer the latent states, plan and make good decisions. Coping with partial observability in general is extremely challenging, as a number of worst-case statistical and computational barriers are known in learning Partially Observable Markov Decision Processes (POMDPs). Motivated by the problem structure in several physical applications, as well as a commonly used technique known as "frame stacking", this paper proposes to study a new subcla…
Many real-world applications of reinforcement learning (RL) require the agent to deal with high-dimensional observations such as those generated from a megapixel camera. Prior work has addressed such problems with representation learning, through which the agent can provably extract endogenous, latent state information from raw observations and subsequently plan efficiently. However, such approaches can fail in the presence of temporally correlated noise in the observations, a phenomenon that is common in practice. We initiate the formal study of latent state discovery in the presence of such …
Seoul National University（首尔大学）；Columbia University（哥伦比亚大学）
We consider a sequential assortment selection problem where the user choice is given by a multinomial logit (MNL) choice model whose parameters are unknown. In each period, the learning agent observes a $d$-dimensional contextual information about the user and the $N$ available items, and offers an assortment of size $K$ to the user, and observes the bandit feedback of the item chosen from the assortment. We propose upper confidence bound based algorithms for this MNL contextual bandit. The first algorithm is a simple and practical method which achieves an $\tilde{\mathcal{O}}(d\sqrt{T})$ regr…
Provable Traffic Rule Compliance in Safe Reinforcement Learning on the Open Sea
Technical University of Munich（慕尼黑工业大学）；TUM School of Computation, Information and Technology（慕尼黑工业大学计算、信息与技术学院）；Munich Center for Machine Learning (MCML)（慕尼黑机器学习中心）
We propose a reinforcement learning algorithm for stationary mean-field games, where the goal is to learn a pair of mean-field state and stationary policy that constitutes the Nash equilibrium. When viewing the mean-field state and the policy as two players, we propose a fictitious play algorithm which alternatively updates the mean-field state and the policy via gradient-descent and proximal policy optimization, respectively. Our algorithm is in stark contrast with previous literature which solves each single-agent reinforcement learning problem induced by the iterates mean-field states to th…
University of Pisa（比萨大学）；University of Chieti-Pescara（基耶蒂-佩斯卡拉大学）
Having access to an exploring restart distribution (the so-called wide coverage assumption) is critical with policy gradient methods. This is due to the fact that, while the objective function is insensitive to updates in unlikely states, the agent may still need improvements in those states in order to reach a nearly optimal payoff. For this reason, wide coverage is used in some form when analyzing theoretical properties of practical policy gradient methods. However, this assumption can be unfeasible in certain environments, for instance when learning is online, or when restarts are possible …
Keep Everyone Happy: Online Fair Division of Numerous Items with Few Copies
Singapore-MIT Alliance for Research and Technology（新加坡-麻省理工联合研究技术联盟）；Faculty of ISEE, Kyushu University（九州大学信息科学与工程学院）；Department of Computer Science, National University of Singapore（新加坡国立大学计算机科学系）
Provable Memory Efficient Self-Play Algorithm for Model-free Reinforcement Learning
College of Information Science and Electronic Engineering（信息科学与电子工程学院）；Zhejiang University（浙江大学）；School of Information and Electronics（信息与电子学院）；Beijing Institute of Technology（北京理工大学）；Institute of Acoustics（声学研究所）；Chinese Academy of Sciences（中国科学院）
University of Science and Technology of China（中国科学技术大学）；The Pennsylvania State University（宾夕法尼亚州立大学）；The Ohio State University（俄亥俄州立大学）
As representation learning becomes a powerful technique to reduce sample complexity in reinforcement learning (RL) in practice, theoretical understanding of its advantage is still limited. In this paper, we theoretically characterize the benefit of representation learning under the low-rank Markov decision process (MDP) model. We first study multitask low-rank RL (as upstream training), where all tasks share a common representation, and propose a new multitask reward-free algorithm called REFUEL. REFUEL learns both the transition kernel and the near-optimal policy for each task, and outputs a …
University of Illinois at Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）；Princeton University（普林斯顿大学）；Northwestern University（西北大学）；Tencent AI Lab（腾讯人工智能实验室）
We consider the problem of \emph{fully decentralized} multi-agent reinforcement learning (MARL), where the agents are located at the nodes of a time-varying communication network. Specifically, we assume that the reward functions of the agents might correspond to different tasks, and are only known to the corresponding agent. Moreover, each agent makes individual decisions based on both the information observed locally and the messages received from its neighbors over the network. Within this setting, the collective goal of the agents is to maximize the globally averaged return over the networ…
MIT（麻省理工学院）；UC Berkeley（加州大学伯克利分校）；Microsoft Research NYC（微软研究院纽约分院）
How do you incentivize self-interested agents to $\textit{explore}$ when they prefer to $\textit{exploit}$? We consider complex exploration problems, where each agent faces the same (but unknown) MDP. In contrast with traditional formulations of reinforcement learning, agents control the choice of policies, whereas an algorithm can only issue recommendations. However, the algorithm controls the flow of information, and can incentivize the agents to explore via information asymmetry. We design an algorithm which explores all reachable states in the MDP. We achieve provable guarantees similar to…
Artificial Intelligence Institute, University of South Carolina（南卡罗来纳大学人工智能研究所）；Computer Science and Engineering, University of Michigan（密歇根大学计算机科学与工程学院）
Multiagent systems can use commitments as the core of a general coordination infrastructure, supporting both cooperative and non-cooperative interactions. Agents whose objectives are aligned, and where one agent can help another achieve greater reward by sacrificing some of its own reward, should choose a cooperative commitment to maximize their joint reward. We present a solution to the problem of how cooperative agents can efficiently find an (approximately) optimal commitment by querying about carefully-selected commitment choices. We prove structural properties of the agents' values as fun…
An important facet of reinforcement learning (RL) has to do with how the agent goes about exploring the environment. Traditional exploration strategies typically focus on efficiency and ignore safety. However, for practical applications, ensuring safety of the agent during exploration is crucial since performing an unsafe action or reaching an unsafe state could result in irreversible damage to the agent. The main challenge of safe exploration is that characterizing the unsafe states and actions is difficult for large continuous state or action spaces and unknown environments. In this paper, w…
INESC-ID/Instituto Superior Técnico（INESC-ID/高等技术学院）；INRIA Bordeaux Sud-Ouest（法国国立信息与自动化研究院波尔多西南分院）
This paper addresses the problem of learning a task from demonstration. We adopt the framework of inverse reinforcement learning, where tasks are represented in the form of a reward function. Our contribution is a novel active learning algorithm that enables the learning agent to query the expert for more informative demonstrations, thus leading to more sample-efficient learning. For this novel algorithm (Generalized Binary Search for Inverse Reinforcement Learning, or GBS-IRL), we provide a theoretical bound on sample complexity and illustrate its applicability on several different tasks. To …
Sample Complexity of Average-Reward Q-Learning: From Single-agent to Federated Reinforcement Learning
CUHK Department of Statistics and Data Science, Chinese University of Hong Kong（中国香港中文大学统计与数据科学系）；CMU Department of Electrical and Computer Engineering, Carnegie Mellon University（卡内基梅隆大学电气与计算机工程系）；CUHK（中国香港中文大学）；CMU（卡内基梅隆大学）；Yale Department of Statistics and Data Science, Yale University（耶鲁大学统计与数据科学系）
Northwestern University（西北大学）；Princeton University（普林斯顿大学）；Georgia Institute of Technology（佐治亚理工学院）
We study discrete-time mean-field Markov games with infinite numbers of agents where each agent aims to minimize its ergodic cost. We consider the setting where the agents have identical linear state transitions and quadratic cost functions, while the aggregated effect of the agents is captured by the population mean of their states, namely, the mean-field state. For such a game, based on the Nash certainty equivalence principle, we provide sufficient conditions for the existence and uniqueness of its Nash equilibrium. Moreover, to find the Nash equilibrium, we propose a mean-field actor-criti…
