Skip to search formSkip to main contentSkip to account menu

Provably Efficient Reinforcement Learning for Discounted MDPs with Feature Mapping

This paper proposes a novel algorithm which makes use of the feature mapping and obtains a first polynomial regret bound, and suggests that the proposed reinforcement learning algorithm is near-optimal up to a $(1-\gamma)^{-0.5}$ factor.
View PDF on arXiv (opens in a new tab)

Logarithmic Regret for Reinforcement Learning with Linear Function Approximation

It is shown that logarithmic regret is attainable under two recently proposed linear MDP assumptions provided that there exists a positive sub-optimality gap for the optimal action-value function.
View PDF on arXiv (opens in a new tab)

Nearly Minimax Optimal Reinforcement Learning for Discounted MDPs

The modelbased algorithm named UCBVI-γ, which is based on the optimism in the face of uncertainty principle and the Bernstein-type bonus, achieves an Õ (√ SAT/(1− γ) ) regret, which suggests that UCB VI-γ is nearly minimax optimal for discounted MDPs.
View PDF on arXiv (opens in a new tab)

Learning Stochastic Shortest Path with Linear Function Approximation

A novel algorithm with Hoeffding-type confidence sets for learning the linear mixture SSP, which provably achieves an near-optimal regret guarantee and proves a lower bound of Ω( dB (cid:63) √ K ) .
View PDF on arXiv (opens in a new tab)

Provably Efficient Representation Learning in Low-rank Markov Decision Processes

A provably efficient algorithm called ReLEX is proposed that can simultaneously learn the representation and perform exploration and will be strictly better in terms of sample efficiency if the function class of representations enjoys a certain mild “coverage” property over the whole state-action space.
View PDF on arXiv (opens in a new tab)

Nearly Minimax Optimal Reinforcement Learning for Linear Markov Decision Processes

This work proposes the first computationally efficient algorithm that achieves the nearly minimax optimal regret and uses a rare-switching policy to update the value function estimator to control the complexity of the estimated value function class.
View PDF on arXiv (opens in a new tab)

Achieving a Fairer Future by Changing the Past

It is shown that algorithms that are informed about the values of future items can get by without any adjustments, whereas uninformed algorithms require Θ(T ) adjustments, and design an uninformed algorithm that requires only O(T ).
View via Publisher (opens in a new tab)

Locally Differentially Private Reinforcement Learning for Linear Mixture Markov Decision Processes

A novel pε, δq-LDP algorithm for learning a class of Markov decision processes (MDPs) dubbed linear mixture MDPs, and obtains local differential privacy guarantees.
View PDF on arXiv (opens in a new tab)

Robust precision attitude tracking of an uncertain rigid spacecraft based on regulation theory

This paper explores regulation theory for the design of robust precision attitude tracking of an uncertain rigid spacecraft with external disturbances. Focusing on the attitude system in terms of…
View on Wiley (opens in a new tab)

Near-optimal Policy Optimization Algorithms for Learning Adversarial Linear Mixture MDPs

This paper proposes an optimistic policy optimization algorithm POWERS and shows that it can achieve regret, and proves a matching lower bound of (cid:101) Ω( dH √ T ) up to logarithmic factors.
View PDF on arXiv (opens in a new tab)
By clicking accept or continuing to use the site, you agree to the terms outlined in our Privacy Policy (opens in a new tab), Terms of Service (opens in a new tab), and Dataset License (opens in a new tab)