Identifying semantic relations in context: Near-misses and overlaps

Alla Rozovskaya, Roxana Girju

Research output: Contribution to conferencePaper

Abstract

This paper addresses the problem of semantic relation identification for a set of relations difficult to differentiate: near-misses and overlaps. Based on empirical observations on a fairly large dataset of such examples we provide an analysis and a taxonomy of such cases. Using this taxonomy we create various contingency sets of relations. These semantic categories are automatically identified by training and testing three state-of-the-art semantic classifiers employing various feature sets. The results show that in order to identify such near-misses and overlaps accurately, a semantic relation identification system needs to go beyond the ontological information of the two nouns and rely heavily on contextual and pragmatic knowledge.

Original languageEnglish (US)
StatePublished - Sep 2009

Fingerprint

Semantics
Taxonomies
Identification (control systems)
Classifiers
Testing

Keywords

  • Lexical semantics
  • Machine learning
  • Semantic relations

ASJC Scopus subject areas

  • Artificial Intelligence
  • Computer Science Applications
  • Software
  • Electrical and Electronic Engineering

Cite this

Identifying semantic relations in context : Near-misses and overlaps. / Rozovskaya, Alla; Girju, Roxana.

2009.

Research output: Contribution to conferencePaper

@conference{2f9619b9009741fbb7a52a0bd84e1439,
title = "Identifying semantic relations in context: Near-misses and overlaps",
abstract = "This paper addresses the problem of semantic relation identification for a set of relations difficult to differentiate: near-misses and overlaps. Based on empirical observations on a fairly large dataset of such examples we provide an analysis and a taxonomy of such cases. Using this taxonomy we create various contingency sets of relations. These semantic categories are automatically identified by training and testing three state-of-the-art semantic classifiers employing various feature sets. The results show that in order to identify such near-misses and overlaps accurately, a semantic relation identification system needs to go beyond the ontological information of the two nouns and rely heavily on contextual and pragmatic knowledge.",
keywords = "Lexical semantics, Machine learning, Semantic relations",
author = "Alla Rozovskaya and Roxana Girju",
year = "2009",
month = "9",
language = "English (US)",

}

TY - CONF

T1 - Identifying semantic relations in context

T2 - Near-misses and overlaps

AU - Rozovskaya, Alla

AU - Girju, Roxana

PY - 2009/9

Y1 - 2009/9

N2 - This paper addresses the problem of semantic relation identification for a set of relations difficult to differentiate: near-misses and overlaps. Based on empirical observations on a fairly large dataset of such examples we provide an analysis and a taxonomy of such cases. Using this taxonomy we create various contingency sets of relations. These semantic categories are automatically identified by training and testing three state-of-the-art semantic classifiers employing various feature sets. The results show that in order to identify such near-misses and overlaps accurately, a semantic relation identification system needs to go beyond the ontological information of the two nouns and rely heavily on contextual and pragmatic knowledge.

AB - This paper addresses the problem of semantic relation identification for a set of relations difficult to differentiate: near-misses and overlaps. Based on empirical observations on a fairly large dataset of such examples we provide an analysis and a taxonomy of such cases. Using this taxonomy we create various contingency sets of relations. These semantic categories are automatically identified by training and testing three state-of-the-art semantic classifiers employing various feature sets. The results show that in order to identify such near-misses and overlaps accurately, a semantic relation identification system needs to go beyond the ontological information of the two nouns and rely heavily on contextual and pragmatic knowledge.

KW - Lexical semantics

KW - Machine learning

KW - Semantic relations

UR - http://www.scopus.com/inward/record.url?scp=84866869204&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84866869204&partnerID=8YFLogxK

M3 - Paper

AN - SCOPUS:84866869204

ER -