The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. ex. Some numerals are expressed as "XNUMX".
Copyrights notice
The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. Copyrights notice
As ontologias são consideradas a solução para a heterogeneidade dos dados na Web Semântica (SW), mas também sofrem com o problema da heterogeneidade, que leva à ambiguidade das informações dos dados. A técnica Ontology Meta-Matching (OMM) é capaz de resolver o problema de heterogeneidade da ontologia através da agregação de várias medidas de similaridade para encontrar as entidades heterogêneas. Inspirado no sucesso do Aprendizado por Reforço (RL) na resolução de problemas complexos de otimização, este trabalho propõe uma técnica OMM baseada em RL para resolver o problema de heterogeneidade de ontologias. Primeiro, propomos um novo framework OMM baseado em RL e, em seguida, uma rede neural chamada rede avaliada é proposta para substituir a tabela Q quando escolhemos a próxima ação do agente, que é capaz de reduzir o consumo de memória e o tempo de computação. . Em seguida, para melhor orientar o treinamento da rede neural e melhorar a precisão do agente RL, estabelecemos um banco de memória para extrair informações de profundidade durante o procedimento de treinamento da rede avaliada, e utilizamos outra rede neural chamada rede alvo para salvar o histórico. parâmetros. O experimento usa o famoso benchmark no domínio de correspondência de ontologias para testar o desempenho de nossa abordagem, e as comparações entre Deep Reinforcement Learning(DRL), RL e sistemas de correspondência de ontologias de última geração mostram que nossa abordagem é capaz de determinar efetivamente alta- alinhamentos de qualidade.
Xingsi XUE
Fujian University of Technology
Yirui HUANG
Fujian University of Technology
Zeqing ZHANG
Xiamen University
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copiar
Xingsi XUE, Yirui HUANG, Zeqing ZHANG, "Deep Reinforcement Learning Based Ontology Meta-Matching Technique" in IEICE TRANSACTIONS on Information,
vol. E106-D, no. 5, pp. 635-643, May 2023, doi: 10.1587/transinf.2022DLP0050.
Abstract: Ontologies are regarded as the solution to data heterogeneity on the Semantic Web (SW), but they also suffer from the heterogeneity problem, which leads to the ambiguity of data information. Ontology Meta-Matching technique (OMM) is able to solve the ontology heterogeneity problem through aggregating various similarity measures to find the heterogeneous entities. Inspired by the success of Reinforcement Learning (RL) in solving complex optimization problems, this work proposes a RL-based OMM technique to address the ontology heterogeneity problem. First, we propose a novel RL-based OMM framework, and then, a neural network that is called evaluated network is proposed to replace the Q table when we choose the next action of the agent, which is able to reduce memory consumption and computing time. After that, to better guide the training of neural network and improve the accuracy of RL agent, we establish a memory bank to mine depth information during the evaluated network's training procedure, and we use another neural network that is called target network to save the historical parameters. The experiment uses the famous benchmark in ontology matching domain to test our approach's performance, and the comparisons among Deep Reinforcement Learning(DRL), RL and state-of-the-art ontology matching systems show that our approach is able to effectively determine high-quality alignments.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2022DLP0050/_p
Copiar
@ARTICLE{e106-d_5_635,
author={Xingsi XUE, Yirui HUANG, Zeqing ZHANG, },
journal={IEICE TRANSACTIONS on Information},
title={Deep Reinforcement Learning Based Ontology Meta-Matching Technique},
year={2023},
volume={E106-D},
number={5},
pages={635-643},
abstract={Ontologies are regarded as the solution to data heterogeneity on the Semantic Web (SW), but they also suffer from the heterogeneity problem, which leads to the ambiguity of data information. Ontology Meta-Matching technique (OMM) is able to solve the ontology heterogeneity problem through aggregating various similarity measures to find the heterogeneous entities. Inspired by the success of Reinforcement Learning (RL) in solving complex optimization problems, this work proposes a RL-based OMM technique to address the ontology heterogeneity problem. First, we propose a novel RL-based OMM framework, and then, a neural network that is called evaluated network is proposed to replace the Q table when we choose the next action of the agent, which is able to reduce memory consumption and computing time. After that, to better guide the training of neural network and improve the accuracy of RL agent, we establish a memory bank to mine depth information during the evaluated network's training procedure, and we use another neural network that is called target network to save the historical parameters. The experiment uses the famous benchmark in ontology matching domain to test our approach's performance, and the comparisons among Deep Reinforcement Learning(DRL), RL and state-of-the-art ontology matching systems show that our approach is able to effectively determine high-quality alignments.},
keywords={},
doi={10.1587/transinf.2022DLP0050},
ISSN={1745-1361},
month={May},}
Copiar
TY - JOUR
TI - Deep Reinforcement Learning Based Ontology Meta-Matching Technique
T2 - IEICE TRANSACTIONS on Information
SP - 635
EP - 643
AU - Xingsi XUE
AU - Yirui HUANG
AU - Zeqing ZHANG
PY - 2023
DO - 10.1587/transinf.2022DLP0050
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E106-D
IS - 5
JA - IEICE TRANSACTIONS on Information
Y1 - May 2023
AB - Ontologies are regarded as the solution to data heterogeneity on the Semantic Web (SW), but they also suffer from the heterogeneity problem, which leads to the ambiguity of data information. Ontology Meta-Matching technique (OMM) is able to solve the ontology heterogeneity problem through aggregating various similarity measures to find the heterogeneous entities. Inspired by the success of Reinforcement Learning (RL) in solving complex optimization problems, this work proposes a RL-based OMM technique to address the ontology heterogeneity problem. First, we propose a novel RL-based OMM framework, and then, a neural network that is called evaluated network is proposed to replace the Q table when we choose the next action of the agent, which is able to reduce memory consumption and computing time. After that, to better guide the training of neural network and improve the accuracy of RL agent, we establish a memory bank to mine depth information during the evaluated network's training procedure, and we use another neural network that is called target network to save the historical parameters. The experiment uses the famous benchmark in ontology matching domain to test our approach's performance, and the comparisons among Deep Reinforcement Learning(DRL), RL and state-of-the-art ontology matching systems show that our approach is able to effectively determine high-quality alignments.
ER -