A Novel Cascade Binary Tagging Framework for Relational Triple Extraction

Новая каскадная бинарная модель разметки для извлечения реляционных троек
Yi Chang, Jianlin Su, Zhepei Wei, Yue Wang, Yuan Tian
2020-01-01

BERT encodercascade binary taggingknowledge graph constructionoverlapping triplesrelational triple extraction
Extracting relational triples from unstructured text is crucial for large-scale knowledge graph construction.However, few existing works excel in solving the overlapping triple problem where multiple relational triples in the same sentence share the same entities.In this work, we introduce a fresh perspective to revisit the relational triple extraction task and propose a novel cascade binary tagging framework (CASREL) derived from a principled problem formulation.Instead of treating relations as discrete labels as in previous works, our new framework models relations as functions that map subjects to objects in a sentence, which naturally handles the overlapping problem.Experiments show that the CAS-REL framework already outperforms state-ofthe-art methods even when its encoder module uses a randomly initialized BERT encoder, showing the power of the new tagging framework.It enjoys further performance boost when employing a pre-trained BERT encoder, outperforming the strongest baseline by 17.5 and 30.2 absolute gain in F1-score on two public datasets NYT and WebNLG, respectively.In-depth analysis on different scenarios of overlapping triples shows that the method delivers consistent performance gain across all these scenarios.The source code and data are released online 1 .
1
Analysis across different overlapping-triple scenarios shows consistent performance gains, and the authors release the source code and data.
2
CASREL introduces a cascade binary tagging framework that models relations as subject-to-object functions rather than discrete labels.
3
CASREL outperforms state-of-the-art methods even with a randomly initialized BERT encoder, demonstrating the effectiveness of its tagging formulation.
4
The framework naturally addresses overlapping relational triples, including cases where multiple triples share entities within a sentence.
5
Using a pre-trained BERT encoder, CASREL exceeds the strongest baseline by 17.5 and 30.2 absolute F1 points on NYT and WebNLG, respectively.

relational triples in unstructured text

overlapping-triple extraction, particularly the handling of multiple triples sharing entities

Publication Details
Publication Date
2020-01-01
Journal
Publisher
ISSN
Cited by
714
Access Type
Author Information
Authors
Yi Chang
Jianlin Su
Zhepei Wei
Yue Wang
Yuan Tian
Explore further
Open the scid.ai AI chat with a ready-made request: it will find papers on a similar topic and help build a literature review.
Find similar papers in the chat →
Make a presentation
100%