Entity extraction is a boring solved problem

Published 1 January 2007Open access

Marc Vilain, Jennifer Su, Suzi Lubar

Citations17

TL;DR

Empirical results that contradict the prevailing opinion that entity extraction is a boring solved problem are presented and an error analysis is given that suggests research challenges for entity extraction that are neither boring nor solved.

Abstract

This paper presents empirical results that contradict the prevailing opinion that entity extraction is a boring solved problem. In particular, we consider data sets that resemble familiar MUC/ACE data, and report surprisingly poor performance for both commercial and research systems. We then give an error analysis that suggests research challenges for entity extraction that are neither boring nor solved.

Keywords

Computer ScienceDecision Sciences

Incorporating non-local information into information extraction systems by Gibbs sampling

3,035 Citations2005Jenny Rose Finkel, Trond Grenager +1 more

By using simulated annealing in place of Viterbi decoding in sequence models such as HMMs, CMMs, and CRFs, it is possible to incorporate non-local structure while preserving tractable inference.

View PDF

Named Entity recognition without gazetteers

431 Citations1999Andrei Mikheev, Marc Moens +1 more

It is shown that, for the text type and task of this competition, it is sufficient to use relatively small gazetteers of well-known names, rather than large gazetters of low-frequency names.

View PDF

Named entity recognition through classifier combination

403 Citations2003Radu Florian, Abe Ittycheriah +2 more

A classifier-combination experimental framework for named entity recognition in which four diverse classifiers (robust linear classifier, maximum entropy, transformation-based learning, and hidden Markov model) are combined under different conditions is presented.

View PDF

Extracting personal names from email

161 Citations2005Einat Minkov, Richard C. Wang +1 more

Two methods for improving performance of person name recognizers for email are presented: email-specific structural features and a recall-enhancing method which exploits name repetition across multiple documents.

View PDF

Composition of conditional random fields for transfer learning

54 Citations2005Charles Sutton, Andrew McCallum

Joint decoding of separately-trained sequence models is performed, preserving uncertainty between the tasks and allowing information from the new task to affect predictions on the old task.

View PDF

Improving name tagging by reference resolution and relation detection

36 Citations2005Heng Ji, Ralph Grishman

This work uses an N-best approach to generate multiple hypotheses and have them re-ranked by subsequent stages of processing to reduce the errors produced by a Chinese name tagger.

View PDF

Syntax-based semi-supervised named entity tagging

23 Citations2005Behrang Mohit, Rebecca Hwa

It is shown that constituency and dependency parsing constraints are both suitable features to extract NEs and train the classifier and the degradation in accuracy on unfamiliar test cases is low, suggesting that the trained classifier generalizes well.

View PDF

Extraction de noms propres ` a partir de textes vari´ es: probl´ ematique et enjeux

2 Citations2001Leila Kosseim, Thierry Poibeau

L’approche utilisée pour adapter un système d’extraction de noms propres développé pour des textes journalistiques à l’analyse de messages électroniques islamique n’arrive pas à des critères rédactionnels stricts.