New method recovers diverse policies from expert data using state-action pair weighting.
problem Recovering diverse policies from expert trajectories.
method Pointwise mutual information weighted behavioral cloning.
result Effective in focusing on state-action pairs most representative of the style.
Improves off-policy evaluation weights for balanced state-action pair distribution.
problem Imbalance in importance sampling weights for off-policy evaluation of contextual bandits.
method Balanced Off-Policy Evaluation (B-OPE) method that minimizes imbalance to desired counterfactual distribution of state-action pairs.
result Experimental evidence shows B-OPE improves offline policy evaluation in both discrete and continuous action spaces.
New findings reveal discount regularization can be seen as a strong prior, leading to poor performance in unevenly sampled data.
problem Discount regularization leads to poor performance in unevenly sampled data.
method Equivalence theorem showing discount regularization as a strong prior, setting regularization parameters locally for individual state-action pairs.
result Discount regularization can be seen as a strong prior, leading to poor performance in unevenly sampled data.
This work tackles model-based RL by optimizing state-action queries to learn policies with minimal data.
problem Expensive state transitions in practical RL problems limit the use of standard RL algorithms.
method Bayesian optimal experimental design to guide selection of state-action queries.
result Data-efficient RL approach that learns optimal policies with up to 1,000x less data.
This paper unifies three regularization methods in batch reinforcement learning.
problem Learning overly-complex models in batch reinforcement learning.
method Unified weighted average transition matrix framework for three regularization methods.
result Empirical evaluation confirms intuitions about regularization methods' performance.
A new method, Count-MORL, improves offline reinforcement learning by using state-action frequency.
problem Improving offline reinforcement learning performance.
method Integrates count-based conservatism into model-based offline reinforcement learning.
result The learned policy is near-optimal and outperforms existing methods.
Backtracking model predicts state-action pairs leading to high-reward states for efficient RL.
problem Efficiently learning from environments where only a few states yield high reward.
method Backtracking model that predicts state-action pairs leading to high-reward states.
result Improves sample efficiency of RL algorithms across various environments and tasks.
GIM learns by analogy to improve reinforcement learning efficiency.
problem Efficient reinforcement learning with limited data.
method GIM infers unknown dynamics from known dynamics based on spectral properties.
result GIM achieves lower sample complexity and computational efficiency.
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.
MineRL dataset tackles reinforcement learning with over 60M annotated state-action pairs.
problem Sample inefficiency in reinforcement learning methods.
method Large-scale, simulator-paired dataset of human demonstrations in Minecraft.
result Demonstrates the scale, diversity, and difficulty of Minecraft tasks.
The Bellman error is a poor proxy for value function accuracy, even with all state-action pairs.
problem The Bellman error is a poor proxy for the accuracy of the value function.
method Study of the Bellman equation as a surrogate objective for value prediction accuracy.
result The magnitude of the Bellman error is only weakly related to the distance to the true value function, even with all state-action pairs.
MBML improves multi-task RL by inferring task identity from state-action pairs.
problem Multi-task batch reinforcement learning with unseen tasks.
method MBML uses triplet loss and relabeling to robustify task inference.
result Significantly faster convergence on unseen tasks compared to random initialization.
The paper improves Q-learning by incorporating pessimism for better sample efficiency.
problem Improving sample efficiency in asynchronous Q-learning with non-i.i.d. data.
method Developed an algorithmic framework that incorporates the principle of pessimism into asynchronous Q-learning, penalizing infrequently-visited state-action pairs based on suitable lower confidence bounds (LCBs).
result Achieved near-optimal sample complexity, providing theoretical support for the use of pessimism in non-i.i.d. data.
The paper learns policies for MDPs from data, even when only some features are relevant.
problem Learning a policy for MDPs from state-action samples with unknown relevant features.
method Uses ℓ 1 \ell_1 ℓ 1 -regularized logistic regression to recover policy parameters. result Establishes bounds on regret in terms of generalization error and Markov chain ergodic coefficient.
Non-stationary reinforcement learning is challenging due to the complexity of updating value functions.
problem Challenges in non-stationary reinforcement learning, especially in updating value functions.
method Proved a worst-case complexity result for modifying reinforcement learning problems.
result Modifying reinforcement learning problems requires an amount of time almost as large as the number of states.
Paper explores state-action equivalence in RL, improving regret bounds.
problem Improving reinforcement learning performance by leveraging state-action equivalence.
method Introduces a notion of similarity between state-action pairs, defines equivalence structure, and presents algorithms for confidence sets.
result Confidence sets improve RL performance, especially in known equivalence structures.
This work compares human feedback methods for reward learning in bandits.
problem Understanding how human feedback affects the performance of reward learning methods.
method Theoretical comparison of human feedback approaches in offline contextual bandits.
result Human bias and uncertainty in feedback modeling impact the theoretical guarantees of reward learning methods.
DualDICE estimates discounted distribution ratios for reinforcement learning datasets.
problem Accurate estimation of discounted stationary distribution ratios for reinforcement learning applications.
method Behavior-agnostic algorithm that avoids importance weights and is theoretically guaranteed.
result Significantly improves off-policy policy evaluation accuracy compared to existing techniques.
Paper tackles covariate shift in offline IL using less proficient behavior data.
problem Mitigating covariate shift in Imitation Learning with limited data coverage.
method Model-based IL from Offline Data (MILO) framework.
result MILO can combat covariate shift even with sub-optimal behavior policy data.
Unified framework for solving MDPs with stochastic mirror descent.
problem Approximately solving infinite-horizon Markov decision processes (MDPs).
method Primal-dual stochastic mirror descent for MDPs with a unified framework.
result Computes ε-optimal policies with expected samples for both average-reward and discounted MDPs.
Paper solves discounted stochastic games with near-optimal time and sample complexity.
problem Solving discounted stochastic two-player games with optimal complexity.
method Generalizes Q-learning to two-player strategy computation, overcoming limitations of existing methods.
result Near-optimal ε ε ε -strategy computation with polylogarithmic factors in 1 − γ 1 - γ 1 − γ and ε − 2 ε^{-2} ε − 2 . Develops a dynamic mean field theory for reinforcement learning.
problem Finite state and action Bayesian reinforcement learning in large state spaces.
method Analogies with statistical physics, interpreting probabilities as couplings and values as spins, solving mean field equations.
result State-action values are statistically independent in the asymptotic state space limit, with exact or approximate equations for computation.
Federated Q-learning achieves linear speedup with heterogeneity, improving sample complexity.
problem Collaborative learning in distributed RL settings with limited data sharing.
method Analyzes synchronous and asynchronous federated Q-learning, proposes importance averaging.
result Achieves linear speedup with heterogeneity, robust to local trajectory heterogeneity.
New algorithm reduces sample complexity for new tasks by leveraging prior knowledge.
problem Designing reinforcement learning agents that reduce sample complexity for new tasks.
method Designing an algorithm that quickly identifies an accurate solution by seeking informative state-action pairs from related tasks, using a generative model.
result PAC bounds on sample complexity demonstrate the benefits of using prior knowledge.
Deep Q-Learning method for Nash equilibria in stochastic games.
problem Model-free learning for multi-agent stochastic games, especially for general-sum games.
method Data-efficient Deep-Q-learning using local linear-quadratic expansion parametrized by deep neural networks.
result The algorithm learns optimal actions for stochastic games without experiencing all state-action pairs.
Proposes Optimistic Pessimistically Initialised Q-Learning (OPIQ) for better exploration in RL.
problem Pessimistic initialisation of Q-values in deep RL leads to poor exploration performance.
method Augments pessimistically initialised Q-values with count-based bonuses to ensure optimism.
result OPIQ outperforms non-optimistic DQN variants in hard exploration tasks.
We seek to learn an effective policy for a Markov Decision Process (MDP) with continuous states via Q-Learning. Given a set of basis functions over state action pairs we search for a corresponding set of linear weights that minimizes the mean Bellman residual. Our algorithm uses a Kalman filter model to estimate those …
Paper improves RL algorithms with known optimal Q-value predictions.
problem Improving RL algorithms with known optimal Q-value predictions.
method Shows how to improve regret bounds by distilling known optimal Q-value predictions.
result Achieves sublinear regret with known predictions, even if not distillation.
State-only imitation learning improves dexterous manipulation learning from videos.
problem High sample complexity in complex domains like dexterous manipulation.
method Train an inverse dynamics model to predict actions from states and train the policy jointly.
result Performs on par with state-action approaches and outperforms RL alone.
Reinforcement learning with trajectory feedback instead of state-action rewards.
problem Frequent feedback not available in practice.
method Extended reinforcement learning algorithms using trajectory feedback for known and unknown transition models.
result Hybrid optimistic-Thompson Sampling algorithm for unknown transition models.
New RL method handles large state-action spaces with complex models.
problem Complex models and large state-action spaces in reinforcement learning.
method π-KRVI, an optimistic modification of least-squares value iteration using kernel ridge regression.
result First order-optimal regret guarantees under general settings, improving over state of the art.
A new framework for offline RL improves policy flexibility and regularity.
problem Lack of environmental interactions in offline RL leads to poor policy performance.
method Proposes a behavior-regularized implicit policy framework with modified policy-matching methods.
result The framework improves policy effectiveness and robustness beyond static datasets.
New method for natural policy gradients converges linearly.
problem Improving natural policy gradient methods for better convergence.
method Fisher-Rao gradient flow applied to state-action distributions.
result Linear convergence rate with geometry-dependent factor.
New method improves deep policy gradient algorithms by learning relative state values.
problem High sample complexity and instability in policy gradient methods.
method Uses a new state-value function approximation based on residual variance.
result Empirical improvement across diverse continuous control tasks and algorithms.
A new RL paradigm reduces state-action-value function approximation inefficiency.
problem Challenges in state-action-value function approximation for RL.
method State Action Separable Reinforcement Learning (sasRL) decouples action space from value function learning.
result sasRL achieves up to 75% better performance than state-of-the-art MDP-based RL algorithms.
DPS uses posterior sampling for preference-based RL, achieving a first regret guarantee.
problem Formal frameworks for preference-based RL with theoretical analysis.
method Preference-based posterior sampling, Bayesian credit assignment.
result First asymptotic Bayesian no-regret rate for preference-based RL.
The study confirms conditions for Q Q Q -learning with persistent exploration.
problem Formulating conditions for Q Q Q -learning with persistent exploration. method Formulated assumptions for Q Q Q -learning with local and global clocks, ensuring persistent exploration. result The Robbins-Monro conditions are confirmed for Q Q Q -learning with persistent exploration. MetaCURL tackles non-stationary MDPs with optimal dynamic regret.
problem Online learning in non-stationary Markov decision processes.
method MetaCURL uses a meta-algorithm with multiple black-box algorithms and a sleeping expert framework.
result Achieves optimal dynamic regret without prior knowledge of MDP changes.
IDAC improves reinforcement learning efficiency by modeling implicit distributions.
problem Improving sample efficiency in reinforcement learning algorithms.
method IDAC uses two DGNs for a distributional critic and a semi-implicit actor to model implicit policy distributions.
result IDAC outperforms state-of-the-art algorithms on OpenAI Gym environments.
New RL method reduces sample complexity for large state-action spaces.
problem Handling large state-action spaces in RL with general Q-functions.
method Nonparametric Q-learning using kernel ridge regression.
result Sample complexity is order optimal with respect to ε and kernel complexity.
Algorithm minimizes regret in RL by evaluating optimal bias function.
problem Minimizing regret in reinforcement learning models.
method Optimism in the Face of Uncertainty (OFU) principle, evaluating optimal bias function.
result Achieves a regret bound of i l d e O ( S A H T ) ilde{O}(\sqrt{SAHT}) i l d e O ( S A H T ) with known upper bound on optimal bias function. New method reduces regret for sparse adversarial SSP problems.
problem Sparse adversarial Stochastic Shortest Path problem.
method Proposed ℓ r \ell_r ℓ r -norm regularizers for adaptive sparsity. result Regret scales with log M \sqrt{\log M} log M instead of log S A \sqrt{\log SA} log S A . Novel framework proves fast RL convergence in continuous spaces.
problem Analyzing stability in continuous state-action RL.
method Introduces a novel framework to analyze stability properties of RL.
result Highlights two key stability properties and demonstrates their satisfaction in RL.
Expands HMRL to multi-target problem considering agent distances.
problem Learning state-action rules for multiple targets.
method Hierarchical Modular Reinforcement Learning with AT field and C4.5.
result Improved reinforcement learning for multi-agent scenarios.
New method estimates state-action stationary distribution for better off-policy policy evaluation.
problem Accurately estimating state-action stationary distribution for off-policy policy evaluation.
method Estimated Mixture Policy (EMP) for state and state-action stationary distribution corrections.
result Empirical validation shows improved accuracy over state-of-the-art methods.
Kernel-UCBVI algorithm balances exploration and exploitation in metric state-action spaces.
problem Exploration-exploitation dilemma in finite-horizon reinforcement learning with metric state-action spaces.
method Kernel-UCBVI, leveraging smoothness and kernel estimators of rewards and transitions.
result First regret bound for kernel-based RL using smoothing kernels, O ( H 3 K 2 d / ( 2 d + 1 ) ) O(H^3 K^{2d/(2d+1)}) O ( H 3 K 2 d / ( 2 d + 1 ) ) . New method finds exact Pareto front for MO-MDPs efficiently.
problem Finding the exact Pareto front for MO-MDPs is challenging.
method Investigates geometric structure, develops efficient algorithm.
result Pareto front is on boundary of convex polytope of deterministic policies.
New framework for efficient query-based imitation learning.
problem Aligning agent policy with human expert behavior without prior knowledge.
method Adversarial reward query with successor representation.
result Significantly outperforms uncertainty-based methods in query efficiency.