<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
</head>
<body bgcolor="#ffffff" text="#000000">
<p class="MsoNormal" style="margin-bottom: 12pt; text-align: center;"
align="center">LDC2008S05<br>
<b>- <a
href="http://www.ldc.upenn.edu/Catalog/CatalogEntry.jsp?catalogId=LDC2008S05">2005
NIST Language Recognition Evaluation</a> -</b><br>
<br>
<b> </b>LDC2008T09<br>
<b>- <a
href="http://www.ldc.upenn.edu/Catalog/CatalogEntry.jsp?catalogId=LDC2008T09">GALE
Phase 1 Arabic Broadcast News Parallel Text - Part 2</a> -<br>
</b></p>
<p class="MsoNormal" style="margin-bottom: 12pt; text-align: center;"
align="center"><br>
<b>The Linguistic Data Consortium (LDC) would like to announce the
availability of two new publications.</b><br>
</p>
<hr size="2" width="100%">
<p class="MsoNormal" style="margin-bottom: 12pt; text-align: center;"
align="center"><b><br>
New
Publications<br>
<br>
</b></p>
<p>(1) <a
href="http://www.ldc.upenn.edu/Catalog/CatalogEntry.jsp?catalogId=LDC2008T09">GALE
Phase 1 Arabic Broadcast News Parallel Text - Part 2</a> is the second
part of
the three-part GALE Phase 1 Arabic Broadcast News Parallel Text, which,
along
with other corpora, was used as training data in year 1 (Phase 1) of
the
DARPA-funded GALE program. The corpus contains transcripts and English
translations of 10.7 hours of Arabic broadcast news programming
selected from
various sources. This corpus does not contain the audio files from
which the
transcripts and translations were generated. <o:p></o:p></p>
<p>The Arabic broadcast news recordings were selected from four sources
and
four different programs. A manual selection procedure was used to
choose
data appropriate for the GALE program, namely, news and conversation
programs
focusing on current events. Stories on topics such as sports,
entertainment
news, and stock market reports were excluded from the data set. Manual
sentence units/segments (SU) annotation was also performed on
a subset of files following LDC's Quick Rich Transcription
specification. Three types of end of sentence SU were identified:
statement SU, question SU, and incomplete SU. </p>
After transcription and SU annotation, they were reformatted into a
human-readable translation format, and the files were then assigned to
professional translators for careful translation. Translators followed
LDC's GALE Translation guidelines, which describe the makeup of the
translation team, the source, data format, the translation data format,
best practices for translating certain linguistic features (such as
names and speech disfluencies), and quality control procedures applied
to completed translations. <br>
<p> <o:p></o:p></p>
<p style="text-align: center;" align="center"><b>*</b><o:p></o:p></p>
<p>(2) The <a
href="http://www.ldc.upenn.edu/Catalog/CatalogEntry.jsp?catalogId=LDC2008S05">2005
NIST Language Recognition Evaluation</a> corpus contains the evaluation
data, portions of the training
data, the evaluation plan, answer keys and scoring script for the 2005
NIST
(National Institute of Standards and Technology) Language Recognition
Evaluation (LRE). The goal of the LRE is to establish the baseline of
current
performance capability for language recognition of conversational
telephone
speech and to lay the groundwork for further research efforts in the
field.
NIST conducted two previous evaluations in <a
href="http://www.nist.gov/speech/tests/lang/1996/LRE96EvalPlan.pdf">1996</a>
and <a
href="http://www.nist.gov/speech/tests/lang/2003/LRE03EvalPlan-v1.pdf">2003</a>.
For the 2005
NIST LRE, the emphasis was on research directed
toward a general base of technology to be ported to various language
recognition tasks with minimum effort and the development of the
ability to
make more difficult discriminations between similar languages and
dialects of
the same language. <o:p></o:p></p>
<p class="MsoNormal">The task evaluated was the detection of a given
target
language or dialect. From a test segment of speech and a target
language or
dialect, the system to be evaluated determined whether the speech was
from the
target language or dialect. The evaluation consisted of speech from the
following languages and dialects: <o:p></o:p></p>
<ul type="disc">
<li class="MsoNormal" style="">English (American) <o:p></o:p></li>
<li class="MsoNormal" style="">English (Indian) <o:p></o:p></li>
<li class="MsoNormal" style="">Hindi <o:p></o:p></li>
<li class="MsoNormal" style="">Japanese <o:p></o:p></li>
<li class="MsoNormal" style="">Korean <o:p></o:p></li>
<li class="MsoNormal" style="">Mandarin (Mainland) <o:p></o:p></li>
<li class="MsoNormal" style="">Mandarin (<st1:country-region><st1:place>Taiwan</st1:place></st1:country-region>)
<o:p></o:p></li>
<li class="MsoNormal" style="">Spanish (Mexican) <o:p></o:p></li>
<li class="MsoNormal" style="">Tamil <o:p></o:p></li>
</ul>
<p>The 2005 NIST Language Recognition Evaluation Plan, which includes a
description of the evaluation tasks, is included with this release.
Further
information regarding this evaluation is also available at the <a
href="http://www.nist.gov/speech/tests/lang/"><span
style="text-decoration: none;">NIST Language Recognition Evaluation</span></a>
website. <o:p></o:p></p>
<p>Each speech file is one side of a telephone conversation . There are
11,106
speech files in sphere (.sph) format for a total of 44.2 hours of
speech. The speech
data was compiled from LDC's CALLFRIEND corpora and from data collected
by
Oregon Health and <st1:place><st1:placename>Science</st1:placename> <st1:placetype>University</st1:placetype></st1:place>.
<o:p></o:p></p>
<p>Each test segment was prepared using an automatic speech activity
detection
algorithm to identify areas and durations of speech. Segments were
chosen to
contain a specified approximate duration of actual speech. The test
segments
contain three nominal durations of speech: 3 seconds, 10 seconds, and
30
seconds. Performance was evaluated separately for test segments of each
duration. Auxiliary information was included in the SPHERE headers to
document
the source file, start time, and duration of all excerpts that were
used to
construct the segment. <br>
</p>
<br>
<hr size="2" width="100%">
<div align="center"><font face="Courier New, Courier, monospace"><small><small><big><br>
Ilya
Ahtaridis<br>
Membership Coordinator</big><br>
<br>
</small>--------------------------------------------------------------------</small><small><br>
</small></font></div>
<div align="center">
<pre class="moz-signature" cols="72"><font
face="Courier New, Courier, monospace">Linguistic Data Consortium Phone: (215) 573-1275
University of Pennsylvania Fax: (215) 573-2175
3600 Market St., Suite 810 <a
class="moz-txt-link-abbreviated" href="mailto:ldc@ldc.upenn.edu">ldc@ldc.upenn.edu</a>
Philadelphia, PA 19104 USA <a
class="moz-txt-link-freetext" href="http://www.ldc.upenn.edu">http://www.ldc.upenn.edu</a></font></pre>
</div>
</body>
</html>