Removing noisy mentions for distant supervision

Ander Intxaurrondo, Mihai Surdeanu, Oier Lopez De Lacalle, Eneko Agirre

Research output: Contribution to journalArticlepeer-review

15 Scopus citations


Relation Extraction methods based on Distant Supervision rely on true tuples to retrieve noisy mentions, which are then used to train traditional supervised relation extraction methods. In this paper we analyze the sources of noise in the mentions, and explore simple methods to filter out noisy mentions. The results show that a combination of mention frequency cut-off, Pointwise Mutual Information and removal of mentions which are far from the feature centroids of relation labels is able to significantly improve the results of two relation extraction models.

Original languageEnglish (US)
Pages (from-to)41-48
Number of pages8
JournalProcesamiento de Lenguaje Natural
StatePublished - Sep 2013


  • Distant Supervision
  • Information Extraction
  • Learning with Noise
  • Relation Extraction

ASJC Scopus subject areas

  • Language and Linguistics
  • Linguistics and Language
  • Computer Science Applications


Dive into the research topics of 'Removing noisy mentions for distant supervision'. Together they form a unique fingerprint.

Cite this