New RL algorithms improve average-reward performance.
problem Improving RL algorithms for average-reward criteria.
method Developed novel algorithms addressing average-reward criterion directly.
result ATRPO significantly outperforms TRPO in challenging MuJuCo environments.
New TD method stabilizes average-reward learning.
problem Stability issues in average-reward TD learning.
method Implicit fixed point update for average-reward TD( λ λ λ ). result Improved numerical stability and broader step-size range.
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.
Faster algorithms for solving multichain MDPs under average-reward criterion.
problem Navigating towards the best connected component in multichain MDPs.
method Developed algorithms to better solve the navigational subproblem, achieving faster convergence rates.
result Improved rates of convergence and sharper complexity measures for multichain MDPs.
This work improves Q-learning for average-reward MDPs, reducing sample and communication complexities in federated settings.
problem Improving sample complexity of Q-learning for average-reward MDPs.
method Simple Q-learning algorithm with carefully chosen parameters for both single-agent and federated scenarios.
result Established first federated Q-learning algorithm for average-reward MDPs with provable efficiency in sample and communication complexities.
Study improves reinforcement learning for stable long-term performance.
problem Distributionally robust average-reward reinforcement learning for stable long-term performance.
method Proposes two algorithms to achieve near-optimal sample complexity.
result Achieves a sample complexity of O ( ∣ S ∣ ∣ A ∣ t m i x 2 ε − 2 ) O(|\mathbf{S}||\mathbf{A}| t_{\mathrm{mix}}^2\varepsilon^{-2}) O ( ∣ S ∣∣ A ∣ t mix 2 ε − 2 ) for estimating optimal policy and robust average reward. Study non-rectangular robust MDPs for average-reward, finding optimal policies and transient values.
problem Non-rectangular robust Markov decision processes under average-reward criterion.
method Proves history-dependent policies are robust-optimal, introduces transient-value framework, constructs epoch-based policy.
result Existence and properties of robust optimal policies, transient value bounds.
New algorithm for average reward learning with bounded hitting time assumption.
problem Minimizing regret in average reward reinforcement learning with bounded hitting time.
method Optimistic Q-learning with a novel L ‾ \overline{L} L operator for bounded hitting time. result Regret bound of i l d e O ( H 5 S A T ) ilde{O}(H^5 S\sqrt{AT}) i l d e O ( H 5 S A T ) for average reward learning. New algorithm uses Whittle index to improve Q-learning for restless bandits.
problem Optimizing decision-making in multiarmed restless bandits with average reward.
method Introduces a novel reinforcement learning algorithm combining Q-learning and Whittle index policy.
result Demonstrates significant computational gains and excellent empirical performance.
Develops first-order methods for average-reward MDPs with strong guarantees.
problem Lack of strong theoretical guarantees for first-order methods in AMDPs.
method Average-reward stochastic policy mirror descent (SPMD) and variance-reduced temporal difference (VRTD) methods.
result Establishes sample complexity results for solving AMDPs.
New algorithm achieves optimal regret in average reward MDPs without prior bias information.
problem Achieving optimal regret in average reward MDPs with computational efficiency and without prior bias information.
method Projective Mitigated Extended Value Iteration (PMEVI) to compute bias-constrained optimal policies efficiently.
result First tractable algorithm with minimax optimal regret of O ~ ( s p ( h ∗ ) S A T ) \widetilde{\mathrm{O}}(\sqrt{\mathrm{sp}(h^*) S A T}) O ( sp ( h ∗ ) S A T ) . This paper introduces a new reward shaping method for average-reward reinforcement learning.
problem Speeding up convergence to an optimal policy in average-reward reinforcement learning tasks.
method Developed a temporal logic-based approach to automatically generate reward shaping functions.
result The optimal policy can be recovered using the proposed reward shaping framework.
A new model-free algorithm achieves near-optimal regret for infinite-horizon MDPs.
problem Model-free reinforcement learning for infinite-horizon average-reward MDPs.
method Exploration Enhanced Q-learning (EE-QL) for weakly communicating MDPs.
result Achieves O ( T ) O(\sqrt{T}) O ( T ) regret bound for general weakly communicating MDPs. New algorithm learns optimal policy for average reward MDPs with sample complexity matching lower bound.
problem Learning optimal policy for average reward in uniformly ergodic MDPs.
method Developed an estimator with sample complexity of O(|S||A|t_{mix}ε^{-2}).
result First algorithm to match lower bound of existing literature.
New offline RL method handles average-reward MDPs with single-policy coverage.
problem Challenges in offline reinforcement learning due to distribution shift and non-uniform coverage.
method Develops an algorithm based on pessimistic discounted value iteration with quantile clipping.
result First fully single-policy sample complexity bound for average-reward offline RL.
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. New algorithm reduces sample complexity for constrained MDPs.
problem Learning policies in constrained average-reward MDPs.
method Model-based algorithm for relaxed and strict feasibility settings.
result Achieves minimax-optimal bounds for constrained MDPs.
New model-free RL algorithm tackles robust average-reward problems with finite sample complexity analysis.
problem Long-term decision-making in environments with varying dynamics.
method Proposes Robust Halpern Iteration (RHI) algorithm based on a black-box sampling oracle and multi-level Monte-Carlo estimator.
result Achieves ε-optimal robust policy with sample complexity of O(1/ε^(2+o(1))) under generative model setting.
Actor-Critic method achieves optimal regret for unichain MDPs.
problem Scalable regret analysis for infinite-horizon average-reward MDPs.
method NAC-B, a Natural Actor-Critic with batching.
result Order-optimal regret of i l d e O ( T ) ilde{O}(\sqrt{T}) i l d e O ( T ) in infinite-horizon average-reward MDPs. UCRL2-VTR achieves nearly optimal regret for learning MDPs with linear function approximation.
problem Learning infinite-horizon average-reward MDPs with linear function approximation.
method UCRL2-VTR algorithm with Bernstein-type bonus.
result Achieves a regret of i l d e O ( d D T ) ilde{O}(d\sqrt{DT}) i l d e O ( d D T ) with matching lower bound. Optimizes learning policies in average-reward MDPs with improved sample complexity.
problem Learning optimal policies in average-reward MDPs with limited samples.
method Reduces to discounted MDPs and uses improved bounds for variance parameters.
result Establishes minimax optimal sample complexity bound of O(SA(H/ε^2))
New algorithm reduces reinforcement learning regret to sqrt(T) without strong dynamics assumptions.
problem Infinite-horizon average-reward reinforcement learning with linear MDPs.
method Approximate by discounted-reward MDPs and apply optimistic value iteration.
result Achieves O(sqrt(T)) regret with polynomial complexity.
Kernel-based function approximation improves reinforcement learning performance.
problem Average reward reinforcement learning in infinite horizon settings.
method Optimistic algorithm based on kernel ridge regression.
result No-regret performance guarantees and confidence intervals for kernel-based predictions.
Logarithmic regret for continuous-time reinforcement learning.
problem Continuous-time Markov decision processes with unknown transition probabilities and holding times.
method Upper confidence reinforcement learning, mean holding time estimation, stochastic comparison of point processes.
result Logarithmic regret bound achieved in finite time.
New algorithm LOOP learns infinite-horizon AMDPs efficiently with function approximation.
problem Learning optimal policies in infinite-horizon AMDPs with function approximation.
method LOOP combines model-based and value-based methods with novel confidence sets and policy updating.
result LOOP achieves sublinear regret bound of i l d e O ( p o l y ( d , s p ( V ∗ ) ) T β ) ilde{\mathcal{O}}(\mathrm{poly}(d, \mathrm{sp}(V^*)) \sqrt{Tβ} ) i l d e O ( poly ( d , sp ( V ∗ )) T β ) . Optimal sample complexity analysis for plug-in approach in average-reward MDPs.
problem Learning optimal policies in average-reward MDPs with a generative model.
method Plug-in approach that constructs a model estimate and computes an optimal policy.
result Optimal sample complexities for the plug-in approach without prior knowledge of problem parameters.
This work analyzes nonexpansive stochastic approximations with Markovian noise, proving convergence in reinforcement learning.
problem Applying stochastic approximation to reinforcement learning settings with nonexpansive operators.
method Investigates nonexpansive stochastic approximations with Markovian noise, providing asymptotic and finite sample analysis.
result First-time proof of convergence for classical tabular average reward temporal difference learning.
Optimizes learning policies in MDPs with weakly communicating structure.
problem Learning optimal policies in weakly communicating MDPs with generative model.
method Span-based approach, reducing to discounted MDPs for analysis.
result First minimax optimal sample complexity bound for weakly communicating MDPs.
Model-free reinforcement learning is known to be memory and computation efficient and more amendable to large scale problems. In this paper, two model-free algorithms are introduced for learning infinite-horizon average-reward Markov Decision Processes (MDPs). The first algorithm reduces the problem to the discounted-r…
New RL method MAC improves performance in sparse reward settings.
problem Slow mixing in large state spaces or sparse rewards.
method Multi-level Monte Carlo Actor-Critic (MAC) algorithm.
result Achieves convergence rate comparable to state-of-the-art AC algorithms.
Improved exploration in factored average-reward MDPs reduces regret.
problem Minimizing regret in unknown Factored Markov Decision Processes (FMDPs).
method DBN-UCRL strategy, inspired by UCRL2, uses Bernstein-type confidence sets for individual elements of the transition function.
result Achieves a regret bound with a leading term strictly improving over existing bounds.
New method reduces sample complexity for robust reinforcement learning.
problem Finite sample analysis in robust reinforcement learning.
method Stochastic approximation framework with controlled bias, using MLMC techniques and geometric truncation.
result Order-optimal sample complexity of i l d e O ( ε − 2 ) ilde{\mathcal{O}}(ε^{-2}) i l d e O ( ε − 2 ) for robust policy evaluation. Algorithm improves RL model selection for repeated games with utility maximization.
problem Optimal policy learning in repeated games with unknown opponent strategy.
method Proposes MRBEAR for average reward RL, applying to utility maximization in repeated games.
result Regret bound shows linear dependence on number of model classes in average reward RL.
Unified meta algorithms estimate various distribution functionals in infinite-armed bandits.
problem Estimating various distribution functionals in infinite-armed bandits.
method Unified meta algorithms for offline and online settings, achieving optimal sample complexities.
result Online estimation offers significant advantage for certain distribution functionals.
New framework improves restless bandit policies for large numbers of arms.
problem Efficiently compute policies for large numbers of arms in restless bandit problems.
method Follow-the-Virtual-Advice framework, converting single-armed policies to N-armed policies.
result Achieves an O(1/\sqrt{N}) optimality gap in both discrete and continuous settings.
Study optimal offline RL with uncertainty sets and distribution shifts.
problem Optimal offline reinforcement learning with limited data.
method Construct uncertainty sets and distribution shifts, solve robust Markov decision process.
result Least conservative estimator for unknown true distribution.
New RL algorithm tackles non-stationary environments with flexible policy updates.
problem Non-stationary reinforcement learning with time-varying rewards and transition probabilities.
method Model-free policy-based algorithm NS-NAC with restart-based exploration and dynamic learning rates.
result Dynamic regret of i l d e O ( ∣ S ∣ 1 / 2 ∣ A ∣ 1 / 2 Δ T 1 / 6 T 5 / 6 ) ilde{\mathscr O}(|S|^{1/2}|A|^{1/2}Δ_T^{1/6}T^{5/6}) i l d e O ( ∣ S ∣ 1/2 ∣ A ∣ 1/2 Δ T 1/6 T 5/6 ) for both algorithms. We develop an off-policy actor-critic algorithm for learning an optimal policy from a training set composed of data from multiple individuals. This algorithm is developed with a view towards its use in mobile health.
This work tackles risk-sensitive deep RL by optimizing policies with variance constraints.
problem Risk and aleatoric uncertainty in deep reinforcement learning.
method Lagrangian and Fenchel dualities to transform the problem into an unconstrained saddle-point policy optimization problem, and an actor-critic algorithm to iteratively update policy, Lagrange multiplier, and Fenchel dual variable.
result The proposed actor-critic algorithm finds a globally optimal policy at a sublinear rate.
New algorithms improve robust reinforcement learning under uncertainty.
problem Robust reinforcement learning in MDPs with contamination.
method Non-asymptotic convergence analysis of Q Q Q -learning and actor-critic methods. result Efficient algorithms learn robust policies with minimal samples.
Reinforcement learning (RL) has traditionally been understood from an episodic perspective; the concept of non-episodic RL, where there is no restart and therefore no reliable recovery, remains elusive. A fundamental question in non-episodic RL is how to measure the performance of a learner and derive algorithms to max…
New RL method improves on standard discounted RL for operations research.
problem Applying RL to operations research problems, especially with non-zero rewards.
method Near-Blackwell-optimal RL algorithm that assesses average reward per step.
result Proves viability on challenging queuing system problems.
New algorithms for learning MDPs with linear approximations in infinite-horizon settings.
problem Learning infinite-horizon average-reward MDPs with linear function approximation.
method Optimism principle, adversarial linear bandits, Natural Policy Gradient.
result Efficient algorithms with optimal or near-optimal regret bounds.
This work analyzes Q Q Q -learning with adaptive stepsizes for finite-time convergence.
problem Finite-time convergence analysis for average-reward Q Q Q -learning with adaptive stepsizes. method Adaptive stepsizes as local clocks, time-inhomogeneous Markovian reformulation, almost-sure time-varying bounds, conditioning arguments, and Markov chain concentration inequalities.
result Convergence rates of i l d e O ( 1 / k ) ilde{\mathcal{O}}(1/k) i l d e O ( 1/ k ) for mean-square and pointwise mean-square convergence. In this paper, we derive a new model of synaptic plasticity, based on recent algorithms for reinforcement learning (in which an agent attempts to learn appropriate actions to maximize its long-term average reward). We show that these direct reinforcement learning algorithms also give locally optimal performance for the…
We derive and analyze learning algorithms for apprenticeship learning, policy evaluation, and policy gradient for average reward criteria. Existing algorithms explicitly require an upper bound on the mixing time. In contrast, we build on ideas from Markov chain theory and derive sampling algorithms that do not require …
New algorithm reduces regret in infinite MDPs with optimal variance-dependent bounds.
problem Infinite horizon MDPs lack optimal algorithms with low regret.
method Developed a UCB-style algorithm for average-reward and γ-regret.
result Achieved optimal variance-dependent regret bounds for both objectives.
New rule reduces exploration regret to logarithmic, improving bad episode handling.
problem Improving exploration regret in average reward MDPs.
method Replacing Doubling Trick with Vanishing Multiplicative rule in EVI-based algorithms.
result Regret is logarithmic under the new rule, significantly better than linear.