Reinforcement learning (RL) is one of the most vibrant research frontiers in machine
learning and has been recently applied to solve a number of challenging problems. In this …

We provide a theoretical framework for Reinforcement Learning with Human Feedback
(RLHF). We show that when the underlying true reward is linear, under both Bradley-Terry …

Reinforcement learning (RL) is typically viewed as the problem of estimating single-step
policies (for model-free RL) or single-step models (for model-based RL), leveraging the …

Large transformer models trained on diverse datasets have shown a remarkable ability to
learn in-context, achieving high few-shot performance on tasks they were not explicitly …

A compelling use case of offline reinforcement learning (RL) is to obtain a policy initialization
from existing datasets followed by fast online fine-tuning with limited interaction. However …

Abstract Model-based reinforcement learning (RL) algorithms, which learn a dynamics
model from logged experience and perform conservative planning under the learned model …

The use of pessimism, when reasoning about datasets lacking exhaustive exploration has
recently gained prominence in offline reinforcement learning. Despite the robustness it adds …

Offline (or batch) reinforcement learning (RL) algorithms seek to learn an optimal policy from
a fixed dataset without active data collection. Based on the composition of the offline dataset …

Abstract We propose Adversarially Trained Actor Critic (ATAC), a new model-free algorithm
for offline reinforcement learning (RL) under insufficient data coverage, based on the …

Sample-efficiency guarantees for offline reinforcement learning (RL) often rely on strong
assumptions on both the function classes (eg, Bellman-completeness) and the data …