Beyond uniform sampling: Offline reinforcement learning with imbalanced datasets
Offline reinforcement learning (RL) enables learning a decision-making policy without
interaction with the environment. This makes it particularly beneficial in situations where …
interaction with the environment. This makes it particularly beneficial in situations where …
Double pessimism is provably efficient for distributionally robust offline reinforcement learning: Generic algorithm and robust partial coverage
We study distributionally robust offline reinforcement learning (RL), which seeks to find an
optimal robust policy purely from an offline dataset that can perform well in perturbed …
optimal robust policy purely from an offline dataset that can perform well in perturbed …
Revisiting the linear-programming framework for offline rl with general function approximation
Offline reinforcement learning (RL) aims to find an optimal policy for sequential decision-
making using a pre-collected dataset, without further interaction with the environment …
making using a pre-collected dataset, without further interaction with the environment …
Policy finetuning in reinforcement learning via design of experiments using offline data
In some applications of reinforcement learning, a dataset of pre-collected experience is
already availablebut it is also possible to acquire some additional online data to help …
already availablebut it is also possible to acquire some additional online data to help …
When is realizability sufficient for off-policy reinforcement learning?
A Zanette - International Conference on Machine Learning, 2023 - proceedings.mlr.press
Understanding when reinforcement learning algorithms can make successful off-policy
predictions—and when the may fail to do so–remains an open problem. Typically, model …
predictions—and when the may fail to do so–remains an open problem. Typically, model …
Importance weighted actor-critic for optimal conservative offline reinforcement learning
Abstract We propose A-Crab (Actor-Critic Regularized by Average Bellman error), a new
practical algorithm for offline reinforcement learning (RL) in complex environments with …
practical algorithm for offline reinforcement learning (RL) in complex environments with …
Policy learning" without''overlap: Pessimism and generalized empirical Bernstein's inequality
This paper studies offline policy learning, which aims at utilizing observations collected a
priori (from either fixed or adaptively evolving behavior policies) to learn the optimal …
priori (from either fixed or adaptively evolving behavior policies) to learn the optimal …
On sample-efficient offline reinforcement learning: Data diversity, posterior sampling and beyond
T Nguyen-Tang, R Arora - Advances in neural information …, 2024 - proceedings.neurips.cc
We seek to understand what facilitates sample-efficient learning from historical datasets for
sequential decision-making, a problem that is popularly known as offline reinforcement …
sequential decision-making, a problem that is popularly known as offline reinforcement …
Offline primal-dual reinforcement learning for linear mdps
Abstract Offline Reinforcement Learning (RL) aims to learn a near-optimal policy from a fixed
dataset of transitions collected by another policy. This problem has attracted a lot of attention …
dataset of transitions collected by another policy. This problem has attracted a lot of attention …
Harnessing density ratios for online reinforcement learning
The theories of offline and online reinforcement learning, despite having evolved in parallel,
have begun to show signs of the possibility for a unification, with algorithms and analysis …
have begun to show signs of the possibility for a unification, with algorithms and analysis …