Tree-KGQA: An Unsupervised Approach for Question Answering Over Knowledge Graphs

Publikation: Beiträge in ZeitschriftenZeitschriftenaufsätzeForschungbegutachtet

Standard

Tree-KGQA: An Unsupervised Approach for Question Answering Over Knowledge Graphs. / Rony, Md Rashad Al Hasan; Chaudhuri, Debanjan; Usbeck, Ricardo et al.
in: IEEE Access, Jahrgang 10, 01.01.2022, S. 50467-50478.

Publikation: Beiträge in ZeitschriftenZeitschriftenaufsätzeForschungbegutachtet

Harvard

APA

Vancouver

Rony MRAH, Chaudhuri D, Usbeck R, Lehmann J. Tree-KGQA: An Unsupervised Approach for Question Answering Over Knowledge Graphs. IEEE Access. 2022 Jan 1;10:50467-50478. doi: 10.1109/ACCESS.2022.3173355

Bibtex

@article{edded7760ad04aa6bd8d060d8d6a5991,
title = "Tree-KGQA: An Unsupervised Approach for Question Answering Over Knowledge Graphs",
abstract = "Most Knowledge Graph-based Question Answering (KGQA) systems rely on training data to reach their optimal performance. However, acquiring training data for supervised systems is both time-consuming and resource-intensive. To address this, in this paper, we propose Tree-KGQA, an unsupervised KGQA system leveraging pre-trained language models and tree-based algorithms. Entity and relation linking are essential components of any KGQA system. We employ several pre-trained language models in the entity linking task to recognize the entities mentioned in the question and obtain the contextual representation for indexing. Furthermore, for relation linking we incorporate a pre-trained language model previously trained for language inference task. Finally, we introduce a novel algorithm for extracting the answer entities from a KG, where we construct a forest of interpretations and introduce tree-walking and tree disambiguation techniques. Our algorithm uses the linked relation and predicts the tree branches that eventually lead to the potential answer entities. The proposed method achieves 4.5% and 7.1% gains in F1 score in entity linking tasks on LC-QuAD 2.0 and LC-QuAD 2.0 (KBpearl) datasets, respectively, and a 5.4% increase in the relation linking task on LC-QuAD 2.0 (KBpearl). The comprehensive evaluations demonstrate that our unsupervised KGQA approach outperforms other supervised state-of-the-art methods on the WebQSP-WD test set (1.4% increase in F1 score)-without training on the target dataset.",
keywords = "Entity linking, Indexing, Information retrieval, Knowledge based systems, Pre-trained language models, Question answering, Relation linking, Informatics, Business informatics",
author = "Rony, {Md Rashad Al Hasan} and Debanjan Chaudhuri and Ricardo Usbeck and Jens Lehmann",
note = "Publisher Copyright: {\textcopyright} 2013 IEEE.",
year = "2022",
month = jan,
day = "1",
doi = "10.1109/ACCESS.2022.3173355",
language = "English",
volume = "10",
pages = "50467--50478",
journal = "IEEE Access",
issn = "2169-3536",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

RIS

TY - JOUR

T1 - Tree-KGQA

T2 - An Unsupervised Approach for Question Answering Over Knowledge Graphs

AU - Rony, Md Rashad Al Hasan

AU - Chaudhuri, Debanjan

AU - Usbeck, Ricardo

AU - Lehmann, Jens

N1 - Publisher Copyright: © 2013 IEEE.

PY - 2022/1/1

Y1 - 2022/1/1

N2 - Most Knowledge Graph-based Question Answering (KGQA) systems rely on training data to reach their optimal performance. However, acquiring training data for supervised systems is both time-consuming and resource-intensive. To address this, in this paper, we propose Tree-KGQA, an unsupervised KGQA system leveraging pre-trained language models and tree-based algorithms. Entity and relation linking are essential components of any KGQA system. We employ several pre-trained language models in the entity linking task to recognize the entities mentioned in the question and obtain the contextual representation for indexing. Furthermore, for relation linking we incorporate a pre-trained language model previously trained for language inference task. Finally, we introduce a novel algorithm for extracting the answer entities from a KG, where we construct a forest of interpretations and introduce tree-walking and tree disambiguation techniques. Our algorithm uses the linked relation and predicts the tree branches that eventually lead to the potential answer entities. The proposed method achieves 4.5% and 7.1% gains in F1 score in entity linking tasks on LC-QuAD 2.0 and LC-QuAD 2.0 (KBpearl) datasets, respectively, and a 5.4% increase in the relation linking task on LC-QuAD 2.0 (KBpearl). The comprehensive evaluations demonstrate that our unsupervised KGQA approach outperforms other supervised state-of-the-art methods on the WebQSP-WD test set (1.4% increase in F1 score)-without training on the target dataset.

AB - Most Knowledge Graph-based Question Answering (KGQA) systems rely on training data to reach their optimal performance. However, acquiring training data for supervised systems is both time-consuming and resource-intensive. To address this, in this paper, we propose Tree-KGQA, an unsupervised KGQA system leveraging pre-trained language models and tree-based algorithms. Entity and relation linking are essential components of any KGQA system. We employ several pre-trained language models in the entity linking task to recognize the entities mentioned in the question and obtain the contextual representation for indexing. Furthermore, for relation linking we incorporate a pre-trained language model previously trained for language inference task. Finally, we introduce a novel algorithm for extracting the answer entities from a KG, where we construct a forest of interpretations and introduce tree-walking and tree disambiguation techniques. Our algorithm uses the linked relation and predicts the tree branches that eventually lead to the potential answer entities. The proposed method achieves 4.5% and 7.1% gains in F1 score in entity linking tasks on LC-QuAD 2.0 and LC-QuAD 2.0 (KBpearl) datasets, respectively, and a 5.4% increase in the relation linking task on LC-QuAD 2.0 (KBpearl). The comprehensive evaluations demonstrate that our unsupervised KGQA approach outperforms other supervised state-of-the-art methods on the WebQSP-WD test set (1.4% increase in F1 score)-without training on the target dataset.

KW - Entity linking

KW - Indexing

KW - Information retrieval

KW - Knowledge based systems

KW - Pre-trained language models

KW - Question answering

KW - Relation linking

KW - Informatics

KW - Business informatics

UR - http://www.scopus.com/inward/record.url?scp=85130844179&partnerID=8YFLogxK

U2 - 10.1109/ACCESS.2022.3173355

DO - 10.1109/ACCESS.2022.3173355

M3 - Journal articles

AN - SCOPUS:85130844179

VL - 10

SP - 50467

EP - 50478

JO - IEEE Access

JF - IEEE Access

SN - 2169-3536

ER -

DOI