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
PY - 2019/11/1
Y1 - 2019/11/1
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 - Entity linking
KW - Dependency parse tree
KW - Cross-sentence distance
KW - Graph Attention Network
UR - https://www.scopus.com/pages/publications/85076995047
U2 - 10.1007/978-3-030-34223-4_33
DO - 10.1007/978-3-030-34223-4_33
M3 - Conference proceeding contribution
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
A2 - Cheng, Reynold
A2 - Mamoulis, Nikos
A2 - Sun, Yizhou
A2 - Huang, Xin
PB - Springer, Springer Nature
CY - Switzerland
T2 - 20th International Conference on Web Information Systems Engineering, WISE 2019
Y2 - 19 January 2020 through 22 January 2020
ER -