For offline reinforcement learning (RL), model-based methods are expected to be data-
efficient as they incorporate dynamics models to generate more data. However, due to …

We present a framework and algorithms to learn controlled dynamics models using neural
stochastic differential equations (SDEs)--SDEs whose drift and diffusion terms are both …

Model-based methods in reinforcement learning offer a promising approach to enhance
data efficiency by facilitating policy exploration within a dynamics model. However …

Offline reinforcement learning (RL) allows learning sequential behavior from fixed datasets.
Since offline datasets do not cover all possible situations, many methods collect additional …

Current approaches to model-based offline Reinforcement Learning (RL) often incorporate
uncertainty-based reward penalization to address the distributional shift problem. While …

We consider the problem of quantifying uncertainty over expected cumulative rewards in
model-based reinforcement learning. In particular, we focus on characterizing the variance …

Despite the recent success of reinforcement learning (RL) in simulated domains and
industrial applications, sample-efficiency remains a fundamental limitation of many model …

How can autonomous systems learn to operate in the wild, ie, complex, dynamic, and
uncertain real-world environments? Despite recent and significant breakthroughs in artificial …