English
Related papers

Related papers: Variance-reduced $Q$-learning is minimax optimal

200 papers

We study risk-sensitive reinforcement learning in finite discounted MDPs, where a generative model of the MDP is assumed to be available. We consider a family or risk measures called the optimized certainty equivalent (OCE), which includes…

Machine Learning · Computer Science 2026-05-22 Oliver Mortensen , Mohammad Sadegh Talebi

We consider a reinforcement learning setting in which the deployment environment is different from the training environment. Applying a robust Markov decision processes formulation, we extend the distributionally robust $Q$-learning…

Machine Learning · Computer Science 2024-08-02 Shengbo Wang , Nian Si , Jose Blanchet , Zhengyuan Zhou

Using deep neural nets as function approximator for reinforcement learning tasks have recently been shown to be very powerful for solving problems approaching real-world complexity. Using these results as a benchmark, we discuss the role…

Machine Learning · Computer Science 2016-01-21 Vincent François-Lavet , Raphael Fonteneau , Damien Ernst

The exploration-exploitation dilemma has been a central challenge in reinforcement learning (RL) with complex model classes. In this paper, we propose a new algorithm, Monotonic Q-Learning with Upper Confidence Bound (MQL-UCB) for RL with…

Machine Learning · Computer Science 2025-10-06 Heyang Zhao , Jiafan He , Quanquan Gu

Q-Learning is a fundamental off-policy reinforcement learning (RL) algorithm that has the objective of approximating action-value functions in order to learn optimal policies. Nonetheless, it has difficulties in reconciling bias with…

Machine Learning · Computer Science 2024-11-22 Mahammad Humayoo

Although Q-learning is one of the most successful algorithms for finding the best action-value function (and thus the optimal policy) in reinforcement learning, its implementation often suffers from large overestimation of Q-function values…

Machine Learning · Computer Science 2020-10-13 Huaqing Xiong , Lin Zhao , Yingbin Liang , Wei Zhang

Watkins' and Dayan's Q-learning is a model-free reinforcement learning algorithm that iteratively refines an estimate for the optimal action-value function of an MDP by stochastically "visiting" many state-ation pairs [Watkins and Dayan,…

Machine Learning · Computer Science 2021-08-09 Matthew T. Regehr , Alex Ayoub

We consider the problem of learning the optimal action-value function in the discounted-reward Markov decision processes (MDPs). We prove a new PAC bound on the sample-complexity of model-based value iteration algorithm in the presence of…

Machine Learning · Computer Science 2012-07-03 Mohammad Gheshlaghi Azar , Remi Munos , Bert Kappen

In this paper, for POMDPs, we provide the convergence of a Q learning algorithm for control policies using a finite history of past observations and control actions, and, consequentially, we establish near optimality of such limit Q…

Machine Learning · Computer Science 2022-10-27 Ali Devran Kara , Serdar Yuksel

Reinforcement learning (RL) for exponential-utility optimization in discounted Markov decision processes (MDPs) lacks principled value-based algorithms. We address this gap in the fixed risk-aversion setting. Building on the Bellman-type…

Machine Learning · Computer Science 2026-05-11 Gugan Thoppe , L. A. Prashanth , Ankur Naskar , Sanjay Bhat

Reinforcement learning studies how an agent should interact with an environment to maximize its cumulative reward. A standard way to study this question abstractly is to ask how many samples an agent needs from the environment to learn an…

Quantum Physics · Physics 2021-12-21 Daochen Wang , Aarthi Sundaram , Robin Kothari , Ashish Kapoor , Martin Roetteler

The objective in this paper is to obtain fast converging reinforcement learning algorithms to approximate solutions to the problem of discounted cost optimal stopping in an irreducible, uniformly ergodic Markov chain, evolving on a compact…

Systems and Control · Computer Science 2019-10-01 Shuhang Chen , Adithya M. Devraj , Ana Bušić , Sean P. Meyn

For infinite-horizon average-cost criterion problems, there exist relatively few rigorous approximation and reinforcement learning results. In this paper, for Markov Decision Processes (MDPs) with standard Borel spaces, (i) we first provide…

Optimization and Control · Mathematics 2024-12-10 Ali Devran Kara , Serdar Yuksel

The recent work by Dong & Yang (2023) showed for misspecified sparse linear bandits, one can obtain an $O\left(\epsilon\right)$-optimal policy using a polynomial number of samples when the sparsity is a constant, where $\epsilon$ is the…

Machine Learning · Computer Science 2024-07-19 Ally Yalei Du , Lin F. Yang , Ruosong Wang

Popular Maximum Entropy Inverse Reinforcement Learning approaches require the computation of expected state visitation frequencies for the optimal policy under an estimate of the reward function. This usually requires intermediate value…

Machine Learning · Computer Science 2020-08-05 Gabriel Kalweit , Maria Huegle , Moritz Werling , Joschka Boedecker

We propose a new simple and natural algorithm for learning the optimal Q-value function of a discounted-cost Markov Decision Process (MDP) when the transition kernels are unknown. Unlike the classical learning algorithms for MDPs, such as…

Optimization and Control · Mathematics 2019-01-31 Dileep Kalathil , Vivek S. Borkar , Rahul Jain

Average-reward reinforcement learning offers a principled framework for long-term decision-making by maximizing the mean reward per time step. Although Q-learning is a widely used model-free algorithm with established sample complexity in…

Machine Learning · Statistics 2026-01-21 Yuchen Jiao , Jiin Woo , Gen Li , Gauri Joshi , Yuejie Chi

Existing studies indicate that momentum ideas in conventional optimization can be used to improve the performance of Q-learning algorithms. However, the finite-sample analysis for momentum-based Q-learning algorithms is only available for…

Machine Learning · Computer Science 2020-07-31 Bowen Weng , Huaqing Xiong , Lin Zhao , Yingbin Liang , Wei Zhang

This paper aims at presenting a new application of information geometry to reinforcement learning focusing on dynamic treatment resumes. In a standard framework of reinforcement learning, a Q-function is defined as the conditional…

Methodology · Statistics 2022-11-17 Shinto Eguchi

The breakthrough of deep Q-Learning on different types of environments revolutionized the algorithmic design of Reinforcement Learning to introduce more stable and robust algorithms, to that end many extensions to deep Q-Learning algorithm…

Machine Learning · Computer Science 2024-04-16 Mohammed Sabry , Amr M. A. Khalifa