We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:


References & Citations

DBLP - CS Bibliography


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Computation and Language

Title: Multi-Scale Feature and Metric Learning for Relation Extraction

Abstract: Existing methods in relation extraction have leveraged the lexical features in the word sequence and the syntactic features in the parse tree. Though effective, the lexical features extracted from the successive word sequence may introduce some noise that has little or no meaningful content. Meanwhile, the syntactic features are usually encoded via graph convolutional networks which have restricted receptive field. To address the above limitations, we propose a multi-scale feature and metric learning framework for relation extraction. Specifically, we first develop a multi-scale convolutional neural network to aggregate the non-successive mainstays in the lexical sequence. We also design a multi-scale graph convolutional network which can increase the receptive field towards specific syntactic roles. Moreover, we present a multi-scale metric learning paradigm to exploit both the feature-level relation between lexical and syntactic features and the sample-level relation between instances with the same or different classes. We conduct extensive experiments on three real world datasets for various types of relation extraction tasks. The results demonstrate that our model significantly outperforms the state-of-the-art approaches.
Subjects: Computation and Language (cs.CL)
Cite as: arXiv:2107.13425 [cs.CL]
  (or arXiv:2107.13425v1 [cs.CL] for this version)

Submission history

From: Mi Zhang [view email]
[v1] Wed, 28 Jul 2021 15:14:36 GMT (1363kb)

Link back to: arXiv, form interface, contact.