Text Relation Extraction Using Sentence-Relation Semantic Similarity

Mohamed Lubani, Shahrul Azman Mohd Noah

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

There is a huge amount of available information stored in unstructured plain text. Relation Extraction (RE) is an important task in the process of converting unstructured resources into machine-readable format. RE is usually considered as a classification problem where a set of features are extracted from the training sentences and thereafter passed to a classifier to predict the relation labels. Existing methods either manually design these features or automatically build them by means of deep neural networks. However, in many cases these features are general and do not accurately reflect the properties of the input sentences. In addition, these features are only built for the input sentences with no regard to the features of the target relations. In this paper, we follow a different approach to perform the RE task. We propose an extended autoencoder model to automatically build vector representations for sentences and relations from their distinctive features. The built vectors are high abstract continuous vector representations (embeddings) where task related features are preserved and noisy irrelevant features are eliminated. Similarity measures are then used to find the sentence-relation semantic similarities using their representations in order to label sentences with the most similar relations. The conducted experiments show that the proposed model is effective in labeling new sentences with their correct semantic relations.

Original languageEnglish
Title of host publicationMulti-disciplinary Trends in Artificial Intelligence - 13th International Conference, MIWAI 2019, Proceedings
EditorsRapeeporn Chamchong, Kok Wai Wong
PublisherSpringer
Pages3-14
Number of pages12
ISBN (Print)9783030337087
DOIs
Publication statusPublished - 1 Jan 2019
Event13th Multi-disciplinary International Conference on Artificial Intelligence, MIWAI 2019 - Kuala Lumpur, Malaysia
Duration: 17 Nov 201919 Nov 2019

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11909 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference13th Multi-disciplinary International Conference on Artificial Intelligence, MIWAI 2019
CountryMalaysia
CityKuala Lumpur
Period17/11/1919/11/19

Fingerprint

Semantic Similarity
Semantics
Labels
Labeling
Classifiers
Text
Experiments
Similarity Measure
Classification Problems
Classifier
Neural Networks
Predict
Resources
Target

Keywords

  • Embeddings
  • Natural language processing
  • Neural networks
  • Relation extraction

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Lubani, M., & Noah, S. A. M. (2019). Text Relation Extraction Using Sentence-Relation Semantic Similarity. In R. Chamchong, & K. W. Wong (Eds.), Multi-disciplinary Trends in Artificial Intelligence - 13th International Conference, MIWAI 2019, Proceedings (pp. 3-14). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11909 LNAI). Springer. https://doi.org/10.1007/978-3-030-33709-4_1

Text Relation Extraction Using Sentence-Relation Semantic Similarity. / Lubani, Mohamed; Noah, Shahrul Azman Mohd.

Multi-disciplinary Trends in Artificial Intelligence - 13th International Conference, MIWAI 2019, Proceedings. ed. / Rapeeporn Chamchong; Kok Wai Wong. Springer, 2019. p. 3-14 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11909 LNAI).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Lubani, M & Noah, SAM 2019, Text Relation Extraction Using Sentence-Relation Semantic Similarity. in R Chamchong & KW Wong (eds), Multi-disciplinary Trends in Artificial Intelligence - 13th International Conference, MIWAI 2019, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11909 LNAI, Springer, pp. 3-14, 13th Multi-disciplinary International Conference on Artificial Intelligence, MIWAI 2019, Kuala Lumpur, Malaysia, 17/11/19. https://doi.org/10.1007/978-3-030-33709-4_1
Lubani M, Noah SAM. Text Relation Extraction Using Sentence-Relation Semantic Similarity. In Chamchong R, Wong KW, editors, Multi-disciplinary Trends in Artificial Intelligence - 13th International Conference, MIWAI 2019, Proceedings. Springer. 2019. p. 3-14. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-030-33709-4_1
Lubani, Mohamed ; Noah, Shahrul Azman Mohd. / Text Relation Extraction Using Sentence-Relation Semantic Similarity. Multi-disciplinary Trends in Artificial Intelligence - 13th International Conference, MIWAI 2019, Proceedings. editor / Rapeeporn Chamchong ; Kok Wai Wong. Springer, 2019. pp. 3-14 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{4b6b7db9714049968f75a8b3821d7dfa,
title = "Text Relation Extraction Using Sentence-Relation Semantic Similarity",
abstract = "There is a huge amount of available information stored in unstructured plain text. Relation Extraction (RE) is an important task in the process of converting unstructured resources into machine-readable format. RE is usually considered as a classification problem where a set of features are extracted from the training sentences and thereafter passed to a classifier to predict the relation labels. Existing methods either manually design these features or automatically build them by means of deep neural networks. However, in many cases these features are general and do not accurately reflect the properties of the input sentences. In addition, these features are only built for the input sentences with no regard to the features of the target relations. In this paper, we follow a different approach to perform the RE task. We propose an extended autoencoder model to automatically build vector representations for sentences and relations from their distinctive features. The built vectors are high abstract continuous vector representations (embeddings) where task related features are preserved and noisy irrelevant features are eliminated. Similarity measures are then used to find the sentence-relation semantic similarities using their representations in order to label sentences with the most similar relations. The conducted experiments show that the proposed model is effective in labeling new sentences with their correct semantic relations.",
keywords = "Embeddings, Natural language processing, Neural networks, Relation extraction",
author = "Mohamed Lubani and Noah, {Shahrul Azman Mohd}",
year = "2019",
month = "1",
day = "1",
doi = "10.1007/978-3-030-33709-4_1",
language = "English",
isbn = "9783030337087",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer",
pages = "3--14",
editor = "Rapeeporn Chamchong and Wong, {Kok Wai}",
booktitle = "Multi-disciplinary Trends in Artificial Intelligence - 13th International Conference, MIWAI 2019, Proceedings",

}

TY - GEN

T1 - Text Relation Extraction Using Sentence-Relation Semantic Similarity

AU - Lubani, Mohamed

AU - Noah, Shahrul Azman Mohd

PY - 2019/1/1

Y1 - 2019/1/1

N2 - There is a huge amount of available information stored in unstructured plain text. Relation Extraction (RE) is an important task in the process of converting unstructured resources into machine-readable format. RE is usually considered as a classification problem where a set of features are extracted from the training sentences and thereafter passed to a classifier to predict the relation labels. Existing methods either manually design these features or automatically build them by means of deep neural networks. However, in many cases these features are general and do not accurately reflect the properties of the input sentences. In addition, these features are only built for the input sentences with no regard to the features of the target relations. In this paper, we follow a different approach to perform the RE task. We propose an extended autoencoder model to automatically build vector representations for sentences and relations from their distinctive features. The built vectors are high abstract continuous vector representations (embeddings) where task related features are preserved and noisy irrelevant features are eliminated. Similarity measures are then used to find the sentence-relation semantic similarities using their representations in order to label sentences with the most similar relations. The conducted experiments show that the proposed model is effective in labeling new sentences with their correct semantic relations.

AB - There is a huge amount of available information stored in unstructured plain text. Relation Extraction (RE) is an important task in the process of converting unstructured resources into machine-readable format. RE is usually considered as a classification problem where a set of features are extracted from the training sentences and thereafter passed to a classifier to predict the relation labels. Existing methods either manually design these features or automatically build them by means of deep neural networks. However, in many cases these features are general and do not accurately reflect the properties of the input sentences. In addition, these features are only built for the input sentences with no regard to the features of the target relations. In this paper, we follow a different approach to perform the RE task. We propose an extended autoencoder model to automatically build vector representations for sentences and relations from their distinctive features. The built vectors are high abstract continuous vector representations (embeddings) where task related features are preserved and noisy irrelevant features are eliminated. Similarity measures are then used to find the sentence-relation semantic similarities using their representations in order to label sentences with the most similar relations. The conducted experiments show that the proposed model is effective in labeling new sentences with their correct semantic relations.

KW - Embeddings

KW - Natural language processing

KW - Neural networks

KW - Relation extraction

UR - http://www.scopus.com/inward/record.url?scp=85076175173&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85076175173&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-33709-4_1

DO - 10.1007/978-3-030-33709-4_1

M3 - Conference contribution

AN - SCOPUS:85076175173

SN - 9783030337087

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 3

EP - 14

BT - Multi-disciplinary Trends in Artificial Intelligence - 13th International Conference, MIWAI 2019, Proceedings

A2 - Chamchong, Rapeeporn

A2 - Wong, Kok Wai

PB - Springer

ER -