English
Related papers

Related papers: Performance Bounds for Policy-Based Average Reward…

200 papers

We consider the problem of off-policy evaluation for reinforcement learning, where the goal is to estimate the expected reward of a target policy $\pi$ using offline data collected by running a logging policy $\mu$. Standard…

Machine Learning · Computer Science 2020-07-09 Ming Yin , Yu-Xiang Wang

We consider a constrained Markov Decision Problem (CMDP) where the goal of an agent is to maximize the expected discounted sum of rewards over an infinite horizon while ensuring that the expected discounted sum of costs exceeds a certain…

Machine Learning · Computer Science 2024-11-01 Washim Uddin Mondal , Vaneet Aggarwal

There has been growing progress on theoretical analyses for provably efficient learning in MDPs with linear function approximation, but much of the existing work has made strong assumptions to enable exploration by conventional exploration…

Machine Learning · Computer Science 2020-10-23 Andrea Zanette , Alessandro Lazaric , Mykel J. Kochenderfer , Emma Brunskill

There have been many recent advances on provably efficient Reinforcement Learning (RL) in problems with rich observation spaces. However, all these works share a strong realizability assumption about the optimal value function of the true…

Machine Learning · Computer Science 2021-06-23 Christoph Dann , Yishay Mansour , Mehryar Mohri , Ayush Sekhari , Karthik Sridharan

Reinforcement Learning with Verifiable Rewards (RLVR) has become a central post-training paradigm for improving the reasoning capabilities of large language models. Yet existing methods share a common blind spot: they optimize policies…

Machine Learning · Computer Science 2026-04-29 Huaiyang Wang , Xiaojie Li , Deqing Wang , Haoyi Zhou , Zixuan Huang , Yaodong Yang , Jianxin Li , Yikun Ban

Value function approximation is important in modern reinforcement learning (RL) problems especially when the state space is (infinitely) large. Despite the importance and wide applicability of value function approximation, its theoretical…

Machine Learning · Computer Science 2023-02-24 Hanlin Zhu , Ruosong Wang , Jason D. Lee

We study the model-based undiscounted reinforcement learning for partially observable Markov decision processes (POMDPs). The oracle we consider is the optimal policy of the POMDP with a known environment in terms of the average reward over…

Machine Learning · Computer Science 2022-07-19 Yi Xiong , Ningyuan Chen , Xuefeng Gao , Xiang Zhou

To date, distributional reinforcement learning (distributional RL) methods have exclusively focused on the discounted setting, where an agent aims to optimize a discounted sum of rewards over time. In this work, we extend distributional RL…

Machine Learning · Computer Science 2026-01-14 Juan Sebastian Rojas , Chi-Guhn Lee

We show two average-reward off-policy control algorithms, Differential Q-learning (Wan, Naik, & Sutton 2021a) and RVI Q-learning (Abounadi Bertsekas & Borkar 2001), converge in weakly communicating MDPs. Weakly communicating MDPs are the…

Machine Learning · Computer Science 2022-11-08 Yi Wan , Richard S. Sutton

We develop a model selection approach to tackle reinforcement learning with adversarial corruption in both transition and reward. For finite-horizon tabular MDPs, without prior knowledge on the total amount of corruption, our algorithm…

Machine Learning · Computer Science 2024-12-31 Chen-Yu Wei , Christoph Dann , Julian Zimmert

We consider reinforcement learning (RL) in episodic Markov decision processes (MDPs) with linear function approximation under drifting environment. Specifically, both the reward and state transition functions can evolve over time but their…

Machine Learning · Computer Science 2024-04-16 Huozhi Zhou , Jinglin Chen , Lav R. Varshney , Ashish Jagmohan

The classical policy gradient method is the theoretical and conceptual foundation of modern policy-based reinforcement learning (RL) algorithms. Most rigorous analyses of such methods, particularly those establishing convergence guarantees,…

Machine Learning · Computer Science 2026-02-11 Jongmin Lee , Ernest K. Ryu

We consider reinforcement learning (RL) in Markov Decision Processes in which an agent repeatedly interacts with an environment that is modeled by a controlled Markov process. At each time step $t$, it earns a reward, and also incurs a…

Machine Learning · Computer Science 2023-03-16 Rahul Singh , Abhishek Gupta , Ness B. Shroff

In recent years, significant attention has been directed towards learning average-reward Markov Decision Processes (MDPs). However, existing algorithms either suffer from sub-optimal regret guarantees or computational inefficiencies. In…

Machine Learning · Computer Science 2024-06-04 Victor Boone , Zihan Zhang

Empirical researchers and decision-makers spanning various domains frequently seek profound insights into the long-term impacts of interventions. While the significance of long-term outcomes is undeniable, an overemphasis on them may…

Machine Learning · Computer Science 2024-09-17 Peng Wu , Ziyu Shen , Feng Xie , Zhongyao Wang , Chunchen Liu , Yan Zeng

We study a common challenge in reinforcement learning for large language models (LLMs): the Zero-Reward Assumption, where non-terminal actions (i.e., intermediate token generations) receive zero task-specific immediate reward, while only…

Machine Learning · Computer Science 2025-06-04 Shenghua He , Tian Xia , Xuan Zhou , Hui Wei

To tackle long planning horizon problems in reinforcement learning with general function approximation, we propose the first algorithm, termed as UCRL-WVTR, that achieves both \emph{horizon-free} and \emph{instance-dependent}, since it…

Machine Learning · Computer Science 2023-12-08 Jiayi Huang , Han Zhong , Liwei Wang , Lin F. Yang

In many real-world applications, it is hard to provide a reward signal in each step of a Reinforcement Learning (RL) process and more natural to give feedback when an episode ends. To this end, we study the recently proposed model of RL…

Machine Learning · Computer Science 2024-05-15 Asaf Cassel , Haipeng Luo , Aviv Rosenberg , Dmitry Sotnikov

State-of-the-art model-based reinforcement learning methods train policies on imagined rollouts. These rollouts are trajectories generated by a learned dynamics model and are scored by a learned reward model, but without querying the true…

Machine Learning · Computer Science 2026-05-13 Nadav Timor , Ravid Shwartz-Ziv , Micah Goldblum , Yann LeCun , David Harel

Policy Iteration (PI) is a classical family of algorithms to compute an optimal policy for any given Markov Decision Problem (MDP). The basic idea in PI is to begin with some initial policy and to repeatedly update the policy to one from an…