Markov decision processes: a tool for sequential decision making under uncertainty
PulseAugur coverage of Markov decision processes: a tool for sequential decision making under uncertainty — every cluster mentioning Markov decision processes: a tool for sequential decision making under uncertainty across labs, papers, and developer communities, ranked by signal.
6 day(s) with sentiment data
-
New frameworks advance Markov Decision Processes for reinforcement learning · 2 sources tracked
Two new arXiv papers introduce advanced frameworks for Markov Decision Processes (MDPs), a key tool in reinforcement learning. The first paper, GRASP-MDP, addresses challenges in offline reinforcement learning by separa…
-
New sub-quadratic method improves bisimulation metric computation for MDPs
Researchers have developed a novel sub-quadratic method for calculating bisimulation metrics in Markov decision processes (MDPs). This new approach utilizes approximate nearest neighbor (ANN) indexing to efficiently sel…
-
New framework optimizes counterfactual policies in stochastic decision-making
Researchers have developed a new method for optimizing counterfactual policies in sequential decision-making scenarios that involve inherent randomness. This approach formalizes counterfactual policy optimization under …
-
New quantum algorithms enhance reinforcement learning in generative models
Researchers have developed new quantum algorithms designed to improve the efficiency of reinforcement learning within generative models. These algorithms leverage quantum subroutines such as quantum mean estimation and …
-
New causal abstraction technique improves MDP scalability
Researchers have developed a new property-driven causal abstraction technique for Markov Decision Processes (MDPs) to address scalability challenges. This method leverages causal relations over state variable predicates…
-
New method simplifies decision trees for Markov decision processes
Researchers have developed a new method called dtControl2+$\\varepsilon$ to create smaller, more understandable decision trees for controllers in Markov decision processes. This approach allows for tunable simplificatio…
-
New dtControl2+$\\varepsilon$ method simplifies decision trees for Markov decision processes
Researchers have developed a new method called dtControl2+$\varepsilon$ to create smaller, more explainable decision trees for Markov decision processes. This technique allows for tunable simplification of controllers b…
-
New algorithms enhance risk-aware reinforcement learning for MDPs
Researchers have developed new online learning algorithms for policy evaluation in Markov decision processes (MDPs) that incorporate dynamic utility-based shortfall risk (UBSR) measures. The proposed UBSR-TD algorithm a…
-
New research offers finite-time convergence guarantees for Natural Policy Gradient algorithms
A new research paper published on arXiv provides the first finite-time convergence guarantees for Natural Policy Gradient (NPG) algorithms in finite-horizon Markov Decision Processes. The study analyzes NPG under both c…
-
Research paper questions robustness of offline RL for treatment recommendations
A new research paper published on arXiv investigates the effectiveness of covariate balance diagnostics in long time horizon Markov decision processes, particularly within the context of offline reinforcement learning f…
-
New PAC learning approach for stochastic games with private info
Researchers have developed a new approach to PAC learning in turn-based stochastic games (TBSGs) with reachability objectives. This work introduces a method that allows for decentralized learning, where players do not s…
-
New framework formalizes risk-aware decision-making in Markov processes
Researchers have introduced risk-aware general-utility Markov decision processes (GUMDPs) to allow agents to optimize risk measures of objective values, enabling a trade-off between expected performance and risk aversio…
-
New algorithm enhances robust reward learning for autonomous agents
Researchers have developed a new machine teaching algorithm designed to improve the robustness of reward learning for autonomous agents. The algorithm operates across multiple Markov Decision Processes (MDPs) and select…
-
Researchers provide spectral analysis and convergence guarantees for dueling Q-learning
This paper presents a spectral analysis of dueling Q-learning, an extension of the Q-learning algorithm used in reinforcement learning. The research focuses on providing theoretical understanding and convergence guarant…
-
New framework enhances digital twins with online Bayesian learning
Researchers have developed a new framework for adaptive digital twins that enhances their value in civil engineering applications. This approach utilizes dynamic Bayesian networks to model the interaction between physic…
-
New framework learns state representations from trajectories without rewards
Researchers have developed a novel state representation framework for Markov decision processes (MDPs) that learns directly from state trajectories without needing reward signals or explicit action data. This method foc…
-
Quantum Bayesian Networks accelerate reinforcement learning in complex environments
Researchers have developed Quantum Bayesian Reinforcement Learning (QBRL), a hybrid quantum-classical algorithm designed to enhance decision-making in partially observable environments. This new approach leverages quant…
-
New method identifies probabilistic causes in uncertain decision-making processes
Researchers have developed a new method for identifying probabilistic causes in Markov Decision Processes (MDPs), which are used for sequential decision-making under uncertainty. This novel approach, detailed in a recen…
-
New method identifies probabilistic causes in Markov Decision Processes
Researchers have developed a new method for identifying probabilistic causes in Markov Decision Processes (MDPs) that offers probabilistic guarantees. This approach addresses limitations in existing methods by focusing …
-
AI policies learn cybersecurity penetration testing faster with history aggregation
Researchers have developed and evaluated reinforcement learning policies for penetration testing in cybersecurity scenarios with partial observability. They compared several Proximal Policy Optimization (PPO) variants, …