New algorithm solves uncertain Markov decision processes using Wasserstein uncertainty.
problem Solving Markov decision processes with uncertain transition probabilities.
method Distributionally robust Q Q Q -learning algorithm for Wasserstein uncertainty. result Convergence of the algorithm proved and demonstrated with real data.
Study optimality in safety-constrained Markov decision processes using asynchronous value iteration and modified Q-learning.
problem Optimality in safety-constrained Markov decision processes with multichain structure.
method Formulated as a zero-sum game, constructed asynchronous value iteration scheme and modified Q-learning algorithm.
result Resolved Bellman's principle of optimality for multichain Markov decision processes and provided learning algorithms.
We optimize saddle-point problems for large-scale Markov decision processes.
problem Optimizing policies in large-scale Markov decision processes.
method Characterized conditions for convergence and designed an optimization algorithm.
result Our algorithm converges faster and is state-space independent.
We develop robust Markov Decision Processes with risk measures for uncertain environments.
problem Uncertainty in Markov Decision Processes and its impact on risk measures.
method Formulation as a Stackelberg game, robust cost and value iterations, existence of optimal policies.
result Existence of deterministic optimal policies for robust optimization and risk measures.
Overview of risk-sensitive Markov decision processes with Optimized Certainty Equivalent.
problem Optimizing decision-making under risk in Markov processes.
method Analyzes risk-sensitive criteria using Optimized Certainty Equivalent, including entropic risk and Conditional Value-at-Risk.
result Conditions for the existence of optimal policies and solution procedures are provided.
A new method for risk-averse decision-making in Markov processes with improved regret bounds.
problem Risk-averse decision-making in Markov processes.
method Introduces mini-batch measures and multipattern risk-averse problems in a feature-based Q Q Q -learning method. result Proves a high-probability regret bound of O ( H 2 N H K ) \mathcal{O}\big(H^2 N^H \sqrt{ K}\big) O ( H 2 N H K ) for the Q Q Q -learning method. New algorithms learn in complex decision-making problems with smooth transitions.
problem Learning in complex decision-making problems with smooth transitions.
method UCB and PSRL philosophies applied to episodic Markov decision processes with kernel approximation.
result Low regret learning achieved in continuous state and action spaces.
Simple method improves exploration in various decision problems.
problem Improving exploration in sequential decision problems.
method Parameterized Exploration (PE) method that considers time horizon and state of knowledge.
result PE outperforms un-tuned methods in various bandit and decision problem settings.
Paper tests Markov assumption in sequential decision making.
problem Testing the Markov assumption in sequential decision making.
method Forward-Backward Learning procedure to test MA without assuming parametric forms.
result The proposed test plays a crucial role in identifying optimal policies in complex decision processes.
Q-learning for average cost MDPs gets a concentration bound.
problem Finding bounds for Q-learning in average cost MDPs.
method Derives a concentration bound using shortest path problem equivalence.
result Numerical comparison with relative value iteration shows the bound's effectiveness.
New algorithm identifies best policy in MDPs faster.
problem Identifying the best policy in Markov Decision Processes.
method Problem-dependent lower bound and first algorithm with instance-specific sample complexity.
result First algorithm with reduced exploration rate for faster convergence.
Paper tackles identifying an odd arm in a multi-armed bandit with restless Markov processes and trembling hand.
problem Identifying an odd arm in a multi-armed bandit with restless Markov processes and trembling hand.
method Derive asymptotic lower bound on expected time to identify the odd arm, stitch together parameterised solutions to MDPs.
result First known asymptotic lower bound on expected time to identify the odd arm, with vanishing error probability.
This paper analyzes risk-sensitive reinforcement learning with Conditional Value-at-Risk (CVaR) for robust Markov Decision Processes.
problem Risk-sensitive reinforcement learning for robust Markov Decision Processes (RMDPs) with state-action-dependent ambiguity sets.
method The paper establishes a connection between robustness and risk sensitivity, defining a new risk measure NCVaR and proposing value iteration algorithms.
result The proposed approach using NCVaR optimization and value iteration algorithms can solve problems with state-action-dependent ambiguity sets.
We address the problem of inverse reinforcement learning in Markov decision processes where the agent is risk-sensitive. In particular, we model risk-sensitivity in a reinforcement learning framework by making use of models of human decision-making having their origins in behavioral psychology, behavioral economics, an…
Develops new reinforcement learning methods for complex constrained decision-making problems.
problem Complex constrained decision-making problems with a continuum of constraints.
method Proposes semi-infinitely constrained Markov decision processes (SICMDPs) and two reinforcement learning algorithms: SI-CRL and SI-CPO.
result Demonstrates the effectiveness of SI-CRL and SI-CPO in solving complex sequential decision-making tasks.
Algorithm finds safe zones in policy Markov Decision Processes to limit trajectory escape.
problem Finding safe zones in policy Markov Decision Processes to limit trajectory escape.
method Bi-criteria approximation learning algorithm with polynomial sample complexity.
result Achieves almost 2 approximation for both escape probability and safe zone size.
Advances in mobile computing technologies have made it possible to monitor and apply data-driven interventions across complex systems in real time. Markov decision processes (MDPs) are the primary model for sequential decision problems with a large or indefinite time horizon. Choosing a representation of the underlying…
New algorithm tackles constrained Markov decision processes with peak constraints.
problem Optimizing dynamic systems with peak constraints.
method Model-free algorithm converting PCMDP to unconstrained problem, applying Q-learning.
result Algorithm achieves ( ε , p ) (ε,p) ( ε , p ) -PAC policy under certain conditions. This paper solves the open problem of computing Bayes optimal prediction for decision trees using a Markov chain Monte Carlo method.
problem Computing the Bayes optimal prediction for decision trees is infeasible due to an infeasible summation over all division patterns of a feature space.
method Solved the open problem using a Markov chain Monte Carlo method with adaptively tuned step size.
result Computed the Bayes optimal prediction for decision trees using a Markov chain Monte Carlo method.
Stochastic domains often involve risk-averse decision makers. While recent work has focused on how to model risk in Markov decision processes using risk measures, it has not addressed the problem of solving large risk-averse formulations. In this paper, we propose and analyze a new method for solving large risk-averse …
The paper tackles batch policy learning in Markov Decision Processes, focusing on average reward maximization.
problem Maximizing long-term average reward in Markov Decision Processes with batch learning.
method Doubly robust estimator for average reward, optimization algorithm for optimal policy, finite-sample regret guarantee.
result The proposed method achieves semiparametric efficiency and provides a finite-sample regret guarantee.
Paper proposes an HMM-based Q-learning for POMDPs.
problem Q-learning struggles with POMDPs due to incomplete state observation.
method Formulates POMDP estimation as HMM estimation, proposing a recursive algorithm to concurrently estimate POMDP parameters and Q function.
result Algorithm converges to optimal Q function and POMDP parameters.
Novel algorithm for Markov decision processes using rank-one approximation.
problem Solving planning and learning problems of Markov decision processes.
method Policy iteration with rank-one approximation of transition probability matrix.
result The proposed algorithm consistently outperforms first-order algorithms and their accelerated versions.
Unified framework connects reinforcement learning and optimal control.
problem Sequential decision-making across different communities.
method Unified modeling framework based on optimizing policies.
result Unified framework includes four universal policy classes.
This paper introduces a new method for optimizing large-scale problems using Markov chain block updates.
problem Optimizing large-scale problems with efficient and natural block selection.
method Markov chain block coordinate descent (BCD) for optimization.
result The method converges for minimizing Lipschitz differentiable functions, with sublinear and linear convergence rates for convex and strongly convex functions, respectively.
The paper tackles robust policy learning in MDPs using statistical methods.
problem Offline data-driven sequential decision making in MDPs.
method Evaluates policies using average rewards centered at policy-induced stationary distributions. Developed a statistically efficient method for estimating robust optimal policies.
result Established a rate-optimal regret bound up to a logarithmic factor.
Sliding window algorithm for RL in non-stationary MDPs with varying rewards and transitions.
problem Reinforcement learning in Markov Decision Processes with changing state-transition probabilities and reward functions.
method Sliding window approach for handling non-stationarity.
result Performance guarantees and optimal window size for the algorithm, along with a sample complexity bound.
Study minimizes risk in MDPs with spectral measures.
problem Minimizing risk in MDPs with spectral measures.
method Splitting into inner and outer minimization problems; solving inner as MDP; proving existence for outer.
result Existence and solution methods for the outer minimization problem.
We consider the inverse reinforcement learning problem, that is, the problem of learning from, and then predicting or mimicking a controller based on state/action data. We propose a statistical model for such data, derived from the structure of a Markov decision process. Adopting a Bayesian approach to inference, we sh…
Decision makers, such as doctors and judges, make crucial decisions such as recommending treatments to patients, and granting bails to defendants on a daily basis. Such decisions typically involve weighting the potential benefits of taking an action against the costs involved. In this work, we aim to automate this task…
New Q-learning method achieves optimal sample complexity for average-reward problems.
problem Challenges in achieving optimal sample complexity for average-reward Q-learning.
method Synchronous and asynchronous Q-learning with a new contraction principle.
result Optimal O ~ ( ε − 2 ) \widetilde{O}(\varepsilon^{-2}) O ( ε − 2 ) sample complexity guarantees. Improved analysis of UCRL2 with empirical Bernstein inequality reduces exploration-exploitation regret.
problem Exploration-exploitation in communicating Markov Decision Processes.
method Analysis of UCRL2 with Empirical Bernstein inequalities (UCRL2B).
result Regret bound of O ~ ( D Γ S A T ) \widetilde{O}(\sqrt{DΓS A T}) O ( D Γ S A T ) for UCRL2B. New algorithm reduces sample complexity for planning in MDPs.
problem Planning in MDPs with unknown transitions.
method MDP-GapE, a trajectory-based MCTS algorithm.
result Proves upper bound on sample complexity in terms of sub-optimality gaps.
This work extends reinforcement learning to handle non-cumulative objectives.
problem Optimizing functions of rewards rather than their sum in decision processes.
method Mapping NCMDPs to standard MDPs for reinforcement learning.
result Reinforcement learning techniques can be applied to NCMDPs.
A new algorithm tackles active exploration in noisy MDPs.
problem Accurately estimating state mean values in MDPs with varying noise levels.
method Introduces a novel learning algorithm to balance exploration and exploitation.
result Active exploration in MDPs can be more challenging than in MAB.
Reinforcement learning improves uplift modeling's accuracy.
problem Directly modeling the incremental impact of treatments on responses.
method Reformulated as a Markov Decision Process (MDP).
result Significant improvement over previous methods in both synthetic and real-world scenarios.
New methods improve temporal difference learning for policy evaluation in Markov decision processes.
problem Improving temporal difference learning for policy evaluation in Markov decision processes.
method Introduced variance-reduced forms of stochastic approximation to achieve non-asymptotic, instance-dependent optimality.
result Temporal difference learning is strictly suboptimal, but variance-reduced forms achieve optimality up to logarithmic factors.
A RL algorithm learns optimal multi-threshold policies for MDPs.
problem Overcoming the curse of dimensionality in MDPs.
method Structure-aware RL algorithm exploiting multi-threshold optimal policies.
result The algorithm converges to the optimal policy asymptotically.
Safe RL approach using Lyapunov functions.
problem Concurrent optimization of performance and safety constraints in RL.
method Lyapunov-based approach for CMDPs, transforming DP and RL algorithms.
result Significant performance improvement in balancing constraints and performance.
Optimizes mobile notifications for multiple objectives using reinforcement learning.
problem Optimizing mobile notification systems for multiple objectives.
method End-to-end offline reinforcement learning with Double Deep Q-network and Conservative Q-learning.
result Demonstrates improved performance and benefits of the proposed approach.
Framework for optimizing portfolios under model uncertainty.
problem Optimizing portfolios in volatile markets considering model uncertainty.
method Dynamic programming and robust optimization for Markov decision processes.
result Robust optimization leads to better portfolio strategies in uncertain market conditions.
Paper eliminates warm-up phase for PO in linear MDPs, achieving optimal regret.
problem Costly warm-up phase in PO algorithms for linear MDPs.
method Simple contraction mechanism replaces warm-up phase.
result Achieves rate-optimal regret with improved dependence on problem parameters.
Paper presents an algorithm for optimal regret in communicating Markov decision processes.
problem Achieving optimal regret in Markov decision processes with a communicating assumption.
method The algorithm explicitly tracks the constant K(M) to learn optimally, balancing exploration, co-exploration, and exploitation.
result The algorithm achieves asymptotically optimal regret K ( M ) log ( T ) + o ( log ( T ) ) K(M) \log(T) + \mathrm{o}(\log(T)) K ( M ) log ( T ) + o ( log ( T )) for communicating Markov decision processes. New algorithms approximate state similarity in large MDPs.
problem Computing exact bisimulation metrics in large MDPs is expensive and impractical.
method Developed a new metric tied to behavior policy, and two algorithms for approximating it.
result Presented algorithms that can approximate bisimulation metrics in large, deterministic MDPs.
Detects spiky corruption in CRMDPs to learn optimal policies.
problem Learning optimal policies in environments with imperfect reward functions.
method Characterized spiky reward corruption, introduced algorithm to detect corrupt states.
result Algorithm can detect corrupt states and learn optimal policies.
Proposes a Riemannian optimization for policy improvement in MDPs.
problem Optimizing policy functions in Markov decision processes (MDPs).
method Riemannian proximal optimization algorithm with Gaussian mixture model (GMM).
result Guaranteed convergence and efficacy demonstrated in preliminary experiments.
New RL theory reduces sample complexity for mixing MDPs.
problem Optimal sample complexity for reinforcement learning in mixing MDPs.
method Regeneration-type ideas to analyze mixing times.
result Optimal sample complexity depends on mixing time, not just discount factor.
New RL algorithm learns optimal policies for MDPs using known structure.
problem Overcoming curse of dimensionality and modeling in MDPs.
method Structure-aware online learning algorithm exploiting known threshold policy.
result Proposed algorithm converges to optimal policy with significant speed improvements.