Markov Decision Process

Markov Decision Processes (MDPs) are mathematical frameworks for modeling sequential decision-making problems under uncertainty, aiming to find optimal policies that maximize cumulative rewards. Current research emphasizes efficient algorithms for solving MDPs, particularly in complex settings like partially observable MDPs (POMDPs) and constrained MDPs (CMDPs), often employing techniques like policy gradient methods, Q-learning, and active inference. These advancements are crucial for improving the design and analysis of autonomous systems, robotics, and other applications requiring intelligent decision-making in dynamic environments, with a growing focus on addressing issues of safety, robustness, and sample efficiency.

Papers

September 12, 2024

Fitted Q-Iteration via Max-Plus-Linear Approximation
Y. Liu, M. A. S. Kolarijani
Markov Decision Process Offline Reinforcement Learning Linear Function Approximation Fitted Q Iteration Max Affine

September 9, 2024

Towards Fast Rates for Federated and Multi-Task Reinforcement Learning
Feng Zhu, Robert W. Heath Jr., Aritra Mitra
Markov Decision Process Policy Gradient Optimal Policy Federated Prompt Cooperation Multi Task Reinforcement Learning Fast Rate Policy Gradient Algorithm

September 7, 2024

Sample and Oracle Efficient Reinforcement Learning for MDPs with Linearly-Realizable Value Functions
Zakaria Mhammedi
Markov Decision Process Near Optimal Policy Plausible Learning Linear Markov Decision Process

September 2, 2024

An Examination of Offline-Trained Encoders in Vision-Based Deep Reinforcement Learning for Autonomous Driving
Shawan Mohammed, Alp Argun, Nicolas Bonnotte, Gerd Ascheid
Deep Reinforcement Learning Autonomous Driving Markov Decision Process Partial Observability Perception Module Exam Document

August 29, 2024

August 25, 2024

Decentralized Stochastic Control in Standard Borel Spaces: Centralized MDP Reductions, Near Optimality of Finite Window Local Information, and Q-Learning
Omar Mrani-Zentar, Serdar Yüksel
Markov Decision Process Q Learning Near Optimality Finite Horizon Information Sharing Banach Space Decentralized Stochastic

August 24, 2024

Rethinking State Disentanglement in Causal Reinforcement Learning
Haiyao Cao, Zhen Zhang, Panpan Cai, Yuhang Liu, Jinan Zou, Ehsan Abbasnejad, Biwei Huang, Mingming Gong, Anton van den Hengel, Javen Qinfeng Shi
Reinforcement Learning Markov Decision Process Causal Reinforcement Learning

August 23, 2024

August 20, 2024

August 19, 2024

August 16, 2024

August 14, 2024

Differentiating Policies for Non-Myopic Bayesian Optimization
Darian Nwankwo, David Bindel
Markov Decision Process Bayesian Optimization Gradient Based Optimization Prior Policy Acquisition Function

August 13, 2024

Markov Decision Process

Papers

Fitted Q-Iteration via Max-Plus-Linear Approximation

Towards Fast Rates for Federated and Multi-Task Reinforcement Learning

Sample and Oracle Efficient Reinforcement Learning for MDPs with Linearly-Realizable Value Functions

An Examination of Offline-Trained Encoders in Vision-Based Deep Reinforcement Learning for Autonomous Driving

Near-Optimal Policy Identification in Robust Constrained Markov Decision Processes via Epigraph Form

On Convergence of Average-Reward Q-Learning in Weakly Communicating Markov Decision Processes

Decentralized Stochastic Control in Standard Borel Spaces: Centralized MDP Reductions, Near Optimality of Finite Window Local Information, and Q-Learning

Rethinking State Disentanglement in Causal Reinforcement Learning

Reduce, Reuse, Recycle: Categories for Compositional Reinforcement Learning

Optimally Solving Simultaneous-Move Dec-POMDPs: The Sequential Central Planning Approach

Hybrid Recurrent Models Support Emergent Descriptions for Hierarchical Planning and Control

An End-to-End Reinforcement Learning Based Approach for Micro-View Order-Dispatching in Ride-Hailing

Efficient Reinforcement Learning in Probabilistic Reward Machines

Enhancing Reinforcement Learning Through Guided Search

Model-based RL as a Minimalist Approach to Horizon-Free and Second-Order Bounds

A Factored MDP Approach To Moving Target Defense With Dynamic Threat Modeling and Cost Efficiency

Differentiating Policies for Non-Myopic Bayesian Optimization

Solving Truly Massive Budgeted Monotonic POMDPs with Oracle-Guided Meta-Reinforcement Learning

Heavy-Ball Momentum Accelerated Actor-Critic With Function Approximation

Variance-Reduced Cascade Q-learning: Algorithms and Sample Complexity