• Home
  • Search
  • A Transfer Approach Using Graph Neural Networks in Deep Reinforcement Learning
  • Cite Icon6
  • https://doi.org/10.1609/aaai.v38i15.29571Copy DOI Icon

A Transfer Approach Using Graph Neural Networks in Deep Reinforcement Learning

Show More
  • Abstract
  • Literature Map
  • References
  • Citations
  • Similar Papers
Abstract

Transfer learning (TL) has shown great potential to improve Reinforcement Learning (RL) efficiency by leveraging prior knowledge in new tasks. However, much of the existing TL research focuses on transferring knowledge between tasks that share the same state-action spaces. Further, transfer from multiple source tasks that have different state-action spaces is more challenging and needs to be solved urgently to improve the generalization and practicality of the method in real-world scenarios. This paper proposes TURRET (Transfer Using gRaph neuRal nETworks), to utilize the generalization capabilities of Graph Neural Networks (GNNs) to facilitate efficient and effective multi-source policy transfer learning in the state-action mismatch setting. TURRET learns a semantic representation by accounting for the intrinsic property of the agent through GNNs, which leads to a unified state embedding space for all tasks. As a result, TURRET achieves more efficient transfer with strong generalization ability between different tasks and can be easily combined with existing Deep RL algorithms. Experimental results show that TURRET significantly outperforms other TL methods on multiple continuous action control tasks, successfully transferring across robots with different state-action spaces.

Similar Papers
  • Research Article
  • Citations4

Multi-source sparse broad transfer learning for parkinson's disease diagnosis via speech.

  • Feb 04, 2025
  • Medical & biological engineering & computing
  • Yuchuan Liu +5
  • Research Article
  • Citations58

A new ensemble deep graph reinforcement learning network for spatio-temporal traffic volume forecasting in a freeway network

  • Jan 29, 2022
  • Digital Signal Processing
  • Pan Shang +5
  • PDF
  • Book Chapter
  • Citations7

Graph Neural Networks and Reinforcement Learning: A Survey

  • Nov 15, 2023
  • Fatemeh Fathinezhad +3
  • Conference Article
  • Citations16

Multi-source Transfer Learning with Ensemble for Financial Time Series Forecasting

  • Dec 01, 2020
  • Qi-Qiao He +2
  • Conference Article
  • Citations16

Applying Graph Neural Network in Deep Reinforcement Learning to Optimize Wireless Network Routing

  • Mar 01, 2022
  • Xiao Xu +2
  • Research Article
  • Citations19

Multi-source transfer learning method for enhancing the deployment of deep reinforcement learning in multi-zone building HVAC control

  • Aug 24, 2024
  • Energy & Buildings
  • Fangli Hou +3
  • Research Article
  • Citations14

Multi-source adversarial transfer learning for ultrasound image segmentation with limited similarity

  • Jul 26, 2023
  • Applied Soft Computing
  • Yifu Zhang +10
  • Conference Article
  • Citations9

Towards Accelerated and Robust Rreinforcement Learning with Transfer Learning

  • Jan 01, 2022
  • Ruiyong Cao
  • PDF
  • Research Article
  • Citations28

An innovative heterogeneous transfer learning framework to enhance the scalability of deep reinforcement learning controllers in buildings with integrated energy systems

  • Feb 20, 2024
  • Building Simulation
  • Davide Coraci +3
  • Research Article

Selective Multi-Source Transfer Learning and Ensemble Learning for Piezoelectric Actuator Feedforward Control

  • Jan 08, 2026
  • Actuators
  • Yaqian Hu +3
  • Research Article
  • Citations4

Multimedia Tasks-Oriented Edge Computing Offloading Scheme Based on Graph Neural Network in Vehicular Networks

  • Jan 01, 2025
  • IEEE Access
  • Yong Huang
  • PDF
  • Research Article
  • Citations6

Offloading Strategy Based on Graph Neural Reinforcement Learning in Mobile Edge Computing

  • Jun 18, 2024
  • Electronics
  • Tao Wang +4
  • Research Article
  • Citations47

Predictive Maintenance Model for IIoT-Based Manufacturing: A Transferable Deep Reinforcement Learning Approach

  • Sep 01, 2022
  • IEEE Internet of Things Journal
  • Kevin Shen Hoong Ong +4
  • Research Article
  • Citations211

A multi-source information transfer learning method with subdomain adaptation for cross-domain fault diagnosis

  • Feb 22, 2022
  • Knowledge-Based Systems
  • Jinghui Tian +3
  • Research Article
  • Citations1

Sample effficient deep reinforcement learning for control

  • Dec 15, 2019
  • Research Repository (Delft University of Technology)
  • Tim De Bruin
Cactus Communications logo

Copyright 2026 Cactus Communications. All rights reserved.