Sample-Efficient Reinforcement Learning with Symmetry-Guided Demonstrations for Robotic Manipulation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Enayati, Amir M. Soufi, Zhang, Zengjie, Gupta, Kashish, Najjaran, Homayoun
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915503489941504
author Enayati, Amir M. Soufi
Zhang, Zengjie
Gupta, Kashish
Najjaran, Homayoun
author_facet Enayati, Amir M. Soufi
Zhang, Zengjie
Gupta, Kashish
Najjaran, Homayoun
contents Reinforcement learning (RL) suffers from low sample efficiency, particularly in high-dimensional continuous state-action spaces of complex robotic manipulation tasks. RL performance can improve by leveraging prior knowledge, even when demonstrations are limited and collected from simplified environments. To address this, we define General Abstract Symmetry (GAS) for aggregating demonstrations from symmetrical abstract partitions of the robot environment. We introduce Demo-EASE, a novel training framework using a dual-buffer architecture that stores both demonstrations and RL-generated experiences. Demo-EASE improves sample efficiency through symmetry-guided demonstrations and behavior cloning, enabling strong initialization and balanced exploration-exploitation. Demo-EASE is compatible with both on-policy and off-policy RL algorithms, supporting various training regimes. We evaluate our framework in three simulation experiments using a Kinova Gen3 robot with joint-space control within PyBullet. Our results show that Demo-EASE significantly accelerates convergence and improves final performance compared to standard RL baselines, demonstrating its potential for efficient real-world robotic manipulation learning.
format Preprint
id arxiv_https___arxiv_org_abs_2304_06055
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Sample-Efficient Reinforcement Learning with Symmetry-Guided Demonstrations for Robotic Manipulation
Enayati, Amir M. Soufi
Zhang, Zengjie
Gupta, Kashish
Najjaran, Homayoun
Robotics
Artificial Intelligence
Reinforcement learning (RL) suffers from low sample efficiency, particularly in high-dimensional continuous state-action spaces of complex robotic manipulation tasks. RL performance can improve by leveraging prior knowledge, even when demonstrations are limited and collected from simplified environments. To address this, we define General Abstract Symmetry (GAS) for aggregating demonstrations from symmetrical abstract partitions of the robot environment. We introduce Demo-EASE, a novel training framework using a dual-buffer architecture that stores both demonstrations and RL-generated experiences. Demo-EASE improves sample efficiency through symmetry-guided demonstrations and behavior cloning, enabling strong initialization and balanced exploration-exploitation. Demo-EASE is compatible with both on-policy and off-policy RL algorithms, supporting various training regimes. We evaluate our framework in three simulation experiments using a Kinova Gen3 robot with joint-space control within PyBullet. Our results show that Demo-EASE significantly accelerates convergence and improves final performance compared to standard RL baselines, demonstrating its potential for efficient real-world robotic manipulation learning.
title Sample-Efficient Reinforcement Learning with Symmetry-Guided Demonstrations for Robotic Manipulation
topic Robotics
Artificial Intelligence
url https://arxiv.org/abs/2304.06055