GRAM: Generalization in Deep RL with a Robust Adaptation Module

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Queeney, James, Cai, Xiaoyi, Schperberg, Alexander, Corcodel, Radu, Benosman, Mouhacine, How, Jonathan P.
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908672212336640
author Queeney, James
Cai, Xiaoyi
Schperberg, Alexander
Corcodel, Radu
Benosman, Mouhacine
How, Jonathan P.
author_facet Queeney, James
Cai, Xiaoyi
Schperberg, Alexander
Corcodel, Radu
Benosman, Mouhacine
How, Jonathan P.
contents The reliable deployment of deep reinforcement learning in real-world settings requires the ability to generalize across a variety of conditions, including both in-distribution scenarios seen during training as well as novel out-of-distribution scenarios. In this work, we present a framework for dynamics generalization in deep reinforcement learning that unifies these two distinct types of generalization within a single architecture. We introduce a robust adaptation module that provides a mechanism for identifying and reacting to both in-distribution and out-of-distribution environment dynamics, along with a joint training pipeline that combines the goals of in-distribution adaptation and out-of-distribution robustness. Our algorithm GRAM achieves strong generalization performance across in-distribution and out-of-distribution scenarios upon deployment, which we demonstrate through extensive simulation and hardware locomotion experiments on a quadruped robot.
format Preprint
id arxiv_https___arxiv_org_abs_2412_04323
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle GRAM: Generalization in Deep RL with a Robust Adaptation Module
Queeney, James
Cai, Xiaoyi
Schperberg, Alexander
Corcodel, Radu
Benosman, Mouhacine
How, Jonathan P.
Machine Learning
Artificial Intelligence
Robotics
The reliable deployment of deep reinforcement learning in real-world settings requires the ability to generalize across a variety of conditions, including both in-distribution scenarios seen during training as well as novel out-of-distribution scenarios. In this work, we present a framework for dynamics generalization in deep reinforcement learning that unifies these two distinct types of generalization within a single architecture. We introduce a robust adaptation module that provides a mechanism for identifying and reacting to both in-distribution and out-of-distribution environment dynamics, along with a joint training pipeline that combines the goals of in-distribution adaptation and out-of-distribution robustness. Our algorithm GRAM achieves strong generalization performance across in-distribution and out-of-distribution scenarios upon deployment, which we demonstrate through extensive simulation and hardware locomotion experiments on a quadruped robot.
title GRAM: Generalization in Deep RL with a Robust Adaptation Module
topic Machine Learning
Artificial Intelligence
Robotics
url https://arxiv.org/abs/2412.04323