<?xml version="1.0" standalone="yes"?>
<Paper uid="H05-1062">
  <Title>Proceedings of Human Language Technology Conference and Conference on Empirical Methods in Natural Language Processing (HLT/EMNLP), pages 491-498, Vancouver, October 2005. c(c)2005 Association for Computational Linguistics Robust Named Entity extraction from large spoken archives</Title>
  <Section position="2" start_page="0" end_page="491" type="intro">
    <SectionTitle>
1 Introduction
</SectionTitle>
    <Paragraph position="0"> Named Entity Recognition (NER) is a crucial step in many Information Extraction (IE) tasks. It has been a specific task in several evaluation programs such as the Message Understanding Conferences (MUC), the Conferences on Natural Language Learning (CoNLL), the DARPA HUB-5 program or more recently the French ESTER Rich Transcription program on Broadcast News data. Most of these conferences have studied the impact of using transcripts generated by an Automatic Speech Recognition (ASR) system rather than written texts. It appears from these studies that unlike other IE tasks, NER performance is greatly affected by the Word Error Rate (WER) of the transcripts processed. To tackle this problem, different ideas have been proposed: modeling explicitly the ASR errors (Palmer and Ostendorf, 2001) or using the ASR system alternate hypotheses found in word lattices (Saraclar and Sproat, 2004). However performance in NER decreases dramatically when processing high WER transcripts like the ones that are obtained with unmatched conditions between the ASR training model and the data to process. This paper investigates this phenomenon in the framework of the NER task of the French Rich Transcription program of Broadcast News ESTER (Gravier et al., 2004).</Paragraph>
    <Paragraph position="1"> Several issues are addressed: * how to jointly optimize the ASR and the NER models ? * what is the impact in term of ASR and NER performance of a temporal mismatch between the corpora used to train and test the models and how can it be recovered by means of meta-data information ? * Can metadata information be used for indexing large spoken archives ?  After a quick overview of related works in IE from speech input, we present the ESTER evaluation program ; then we introduce a NER system tightly integrated to the ASR process and show how it can successfully index high WER spoken databases thanks to metadata.</Paragraph>
  </Section>
class="xml-element"></Paper>