Utilising Knowledge Graph Embeddings for Data-to-Text Generation

MIHAEL ARCAN; Nivranshu Pasricha; Peter Paul Buitelaar

doi:http://hdl.handle.net/10379/16526

Utilising Knowledge Graph Embeddings for Data-to-Text Generation

MIHAEL ARCAN, Nivranshu Pasricha, Peter Paul Buitelaar

Research output: Chapter in Book or Conference Publication/Proceeding › Conference Publication › peer-review

Abstract

Data-to-text generation has recently seen a move away from modular and pipeline architectures towards end-to-end architectures based on neural networks. In this work, we employ knowledge graph embeddings and explore their utility for end-to-end approaches in a data-to-text generation task. Our experiments show that using knowledge graph embeddings can yield an improvement of up to 23 BLEU points for seen categories on the WebNLG corpus without modifying the underlying neural network architecture.

Original language	English (Ireland)
Title of host publication	3rd WebNLG Workshop on Natural Language Generation from the Semantic Web (WebNLG+ 2020)
Place of Publication	Dublin, Ireland (Virtual)
DOIs	https://doi.org/http://hdl.handle.net/10379/16526
Publication status	Published - 1 Dec 2020

Authors (Note for portal: view the doc link for the full list of authors)

Authors
Pasricha N; Arcan M; Buitelaar P

Access to Document

http://hdl.handle.net/10379/16526

Cite this

@inproceedings{493c3c35bd0c4c788c308dd64cb26cf9,

title = "Utilising Knowledge Graph Embeddings for Data-to-Text Generation",

abstract = "Data-to-text generation has recently seen a move away from modular and pipeline architectures towards end-to-end architectures based on neural networks. In this work, we employ knowledge graph embeddings and explore their utility for end-to-end approaches in a data-to-text generation task. Our experiments show that using knowledge graph embeddings can yield an improvement of up to 23 BLEU points for seen categories on the WebNLG corpus without modifying the underlying neural network architecture.",

author = "MIHAEL ARCAN and Nivranshu Pasricha and Buitelaar, {Peter Paul}",

year = "2020",

month = dec,

day = "1",

doi = "http://hdl.handle.net/10379/16526",

language = "English (Ireland)",

booktitle = "3rd WebNLG Workshop on Natural Language Generation from the Semantic Web (WebNLG+ 2020)",

}

TY - GEN

T1 - Utilising Knowledge Graph Embeddings for Data-to-Text Generation

AU - ARCAN, MIHAEL

AU - Pasricha, Nivranshu

AU - Buitelaar, Peter Paul

PY - 2020/12/1

Y1 - 2020/12/1

N2 - Data-to-text generation has recently seen a move away from modular and pipeline architectures towards end-to-end architectures based on neural networks. In this work, we employ knowledge graph embeddings and explore their utility for end-to-end approaches in a data-to-text generation task. Our experiments show that using knowledge graph embeddings can yield an improvement of up to 23 BLEU points for seen categories on the WebNLG corpus without modifying the underlying neural network architecture.

AB - Data-to-text generation has recently seen a move away from modular and pipeline architectures towards end-to-end architectures based on neural networks. In this work, we employ knowledge graph embeddings and explore their utility for end-to-end approaches in a data-to-text generation task. Our experiments show that using knowledge graph embeddings can yield an improvement of up to 23 BLEU points for seen categories on the WebNLG corpus without modifying the underlying neural network architecture.

U2 - http://hdl.handle.net/10379/16526

DO - http://hdl.handle.net/10379/16526

M3 - Conference Publication

BT - 3rd WebNLG Workshop on Natural Language Generation from the Semantic Web (WebNLG+ 2020)

CY - Dublin, Ireland (Virtual)

ER -

Utilising Knowledge Graph Embeddings for Data-to-Text Generation

Abstract

Authors (Note for portal: view the doc link for the full list of authors)

Access to Document

Fingerprint

Cite this