TY - GEN
T1 - Entity Disambiguation Based on Parse Tree Neighbours on Graph Attention Network
AU - Xin, Kexuan
AU - Hua, Wen
AU - Liu, Yu
AU - Zhou, Xiaofang
N1 - Publisher Copyright:
© 2019, Springer Nature Switzerland AG.
PY - 2019
Y1 - 2019
N2 - Entity disambiguation (ED) aims to link textual mentions in a document to the correct named entities in a knowledge base (KB). Although global ED model usually outperforms local model by collectively linking mentions based on the topical coherence assumption, it may still incur incorrect entity assignment when a document contains multiple topics. Therefore, we propose to extract global features locally, i.e., among a limited number of neighbouring mentions, to combine the respective superiority of both models. In particular, we derive mention neighbours according to the syntactic distance on a dependency parse tree, and propose a tree connection method CoSimTC to measure the cross-tree distance between mentions. Besides, we extend the Graph Attention Network (GAT) to integrate both local and global features to produce a discriminative representation for each candidate entity. Our experimental results on five widely-adopted public datasets demonstrate better performance compared with state-of-the-art approaches.
AB - Entity disambiguation (ED) aims to link textual mentions in a document to the correct named entities in a knowledge base (KB). Although global ED model usually outperforms local model by collectively linking mentions based on the topical coherence assumption, it may still incur incorrect entity assignment when a document contains multiple topics. Therefore, we propose to extract global features locally, i.e., among a limited number of neighbouring mentions, to combine the respective superiority of both models. In particular, we derive mention neighbours according to the syntactic distance on a dependency parse tree, and propose a tree connection method CoSimTC to measure the cross-tree distance between mentions. Besides, we extend the Graph Attention Network (GAT) to integrate both local and global features to produce a discriminative representation for each candidate entity. Our experimental results on five widely-adopted public datasets demonstrate better performance compared with state-of-the-art approaches.
KW - Cross-sentence distance
KW - Dependency parse tree
KW - Entity linking
KW - Graph Attention Network
UR - http://www.scopus.com/inward/record.url?scp=85076995047&partnerID=8YFLogxK
U2 - 10.1007/978-3-030-34223-4_33
DO - 10.1007/978-3-030-34223-4_33
M3 - Conference article published in proceeding or book
AN - SCOPUS:85076995047
SN - 9783030342227
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 523
EP - 537
BT - Web Information Systems Engineering – WISE 2019 - 20th International Conference, Proceedings
A2 - Cheng, Reynold
A2 - Mamoulis, Nikos
A2 - Sun, Yizhou
A2 - Huang, Xin
PB - Springer
T2 - 20th International Conference on Web Information Systems Engineering, WISE 2019
Y2 - 26 November 2019 through 30 November 2019
ER -