Named Entity Extraction from Broadcast News (1999)
| Venue: | In Proceedings of the DARPA Broadcast News Workshop |
| Citations: | 32 - 0 self |
BibTeX
@INPROCEEDINGS{Miller99namedentity,
author = {David Miller and Richard Schwartz and Ralph Weischedel and Rebecca Stone},
title = {Named Entity Extraction from Broadcast News},
booktitle = {In Proceedings of the DARPA Broadcast News Workshop},
year = {1999},
pages = {37--40}
}
Years of Citing Articles
Bookmark
 |
 |
 |
 |
 |
 |
OpenURL
Abstract
In this paper, we contrast the two tasks of named entity extraction from speech and text both qualitatively and quantitatively in the context of the DARPA 1998 Hub4e-IE evaluation. We will present some top level observations and a detailed engineering analysis of our system 's failures and successes. We explore the effects of word error rate, loss of textual clues, amount of training data, changes in guidelines, and out-of-vocabulary errors. 1. Introduction BBN used the IdentiFinder(tm) system (described in [1]) to perform the Named-Entity extraction spoke of the 1998 Hub4 DARPA evaluation. We annotated with named-entity markup the 175 hours of Broadcast News acoustic modelling data, and trained IdentiFinder's statistical models on it. In test, IdentiFinder segmented the input text into paragraphs by splitting at story boundaries (when working from text) or at 1 second silences (when working from speech). Table 1 shows our official evaluation results for all five transcript conditio...
Citations
No citations identified.
