H2O+: An Improved Framework for Hybrid Offline-and-Online RL with Dynamics Gaps

TL;DR

H2O+ is a new hybrid RL framework that effectively combines offline data and imperfect simulations to improve real-world task performance despite dynamics gaps.

Contribution

The paper introduces H2O+, a flexible algorithm that bridges offline and online RL methods while addressing dynamics gaps between simulation and reality.

Findings

H2O+ outperforms existing cross-domain RL algorithms in simulations.

H2O+ demonstrates superior real-world robotics performance.

The framework offers high flexibility in combining offline and online learning.

Abstract

Solving real-world complex tasks using reinforcement learning (RL) without high-fidelity simulation environments or large amounts of offline data can be quite challenging. Online RL agents trained in imperfect simulation environments can suffer from severe sim-to-real issues. Offline RL approaches although bypass the need for simulators, often pose demanding requirements on the size and quality of the offline datasets. The recently emerged hybrid offline-and-online RL provides an attractive framework that enables joint use of limited offline data and imperfect simulator for transferable policy learning. In this paper, we develop a new algorithm, called H2O+, which offers great flexibility to bridge various choices of offline and online learning methods, while also accounting for dynamics gaps between the real and simulation environment. Through extensive simulation and real-world robotics experiments, we demonstrate superior performance and flexibility over advanced cross-domain online and offline RL algorithms.