2008 CoNLL Shared Task Data

<h3>Introduction</h3><br>
<p>2008 CoNLL Shared Task Data, Linguistic Data Consortium (LDC) catalog number LDC2009T12 and isbn 1-58563-505-7, contains the the trial corpus, training corpus, development and test data for the <a href="http://www.conll.org/previous-tasks" rel="nofollow">2008 CoNLL (Conference on Computational Natural Language Learning) Shared Task Evaluation</a>. The 2008 Shared Task developed syntactic dependency annotations, including information such as named-entity boundaries and the semantic dependencies model roles of both verbal and nominal predicates. The materials in the Shared Task data consist of excerpts from the following corpora: <a href="http://catalog.ldc.upenn.edu/LDC99T42" rel="nofollow">Treebank-3</a> LDC99T42, <a href="http://catalog.ldc.upenn.edu/LDC2005T33" rel="nofollow">BBN Pronoun Coreference and Entity Type Corpus</a> LDC2005T33, <a href="http://catalog.ldc.upenn.edu/LDC2004T14" rel="nofollow">Proposition Bank I</a> LDC2004T14 (PropBank) and <a href="http://catalog.ldc.upenn.edu/LDC2008T23" rel="nofollow">NomBank v 1.0</a> LDC2008T23.</p><br>
<p>The <a href="http://www.conll.org/previous-tasks" rel="nofollow">Conference on Computational Natural Language Learning (CoNLL)</a> is accompanied every year by a shared task intended to promote natural language processing applications and evaluate them in a standard setting. The 2004 and 2005 CoNLL shared tasks were dedicated to semantic role labeling (SRL) in a monolingual setting (English). In 2006 and 2007, the shared tasks were devoted to the parsing of syntactic dependencies and used corpora from up to thirteen languages. The 2008 shared task employed a unified dependency-based formalism and merged the task of syntactic dependency parsing and the task of identifying semantic arguments and labeling them with semantic roles.</p><br>
<p>LDC has also released the following CoNLL Shared Task data sets:</p><br>
<ul><br>
<li>2006 CoNLL Shared Task - Ten Languages (<a href="../../../LDC2015T11">LDC2015T11</a>)</li><br>
<li>2006 CoNLL Shared Task - Arabic &amp; Czech (<a href="../../../LDC2015T12">LDC2015T12</a>)</li><br>
<li>2009 CoNLL Shared Task Part 1 (<a href="../../../LDC2012T03">LDC2012T03</a>)</li><br>
<li>2009 CoNLL Shared Task Part 2 (<a href="../../../LDC2012T04">LDC2012T04</a>)</li><br>
<li>2015-2016 CoNLL Shared Task (<a href="../../../LDC2017T13">LDC2017T13</a>)</li><br>
</ul><br>
<h3>Data</h3><br>
<p>The 2008 shared task was divided into three subtasks:</p><br>
<ol><br>
<li>parsing syntactic dependencies</li><br>
<li>identification and disambiguation of semantic predicates</li><br>
<li>identification of arguments and assignment of semantic roles for each predicate</li><br>
</ol><br>
<p>Several objectives were addressed in this shared task:</p><br>
<ul><br>
<li>SRL was performed and evaluated using a dependency-based representation for both syntactic and semantic dependencies. While SRL on top of a dependency treebank has been addressed before, the approach of the 2008 Shared Task was characterized by the following novelties:<br>
<ol><br>
<li>The constituent-to-dependency conversion strategy transformed all annotated semantic arguments in PropBank and NomBank v 1.0, not just a subset;</li><br>
<li>The annotations addressed propositions centered around both verbal (PropBank) and nominal (NomBank) predicates.</li><br>
</ol><br>
</li><br>
<li>Based on the observation that a richer set of syntactic dependencies improves semantic processing, the syntactic dependencies modeled are more complex than the ones used in the previous CoNLL shared tasks. For example, the corpus includes apposition links, dependencies derived from named entity (NE) structures, and better modeling of long-distance grammatical relations.</li><br>
<li>A practical framework is provided for the joint learning of syntactic and semantic dependencies.</li><br>
</ul><br>
<p>Due to the complexity of the 2008 shared task, only a single language, English, was used.</p><br>
<h3>Samples</h3><br>
<p>An example of the shared task annotations is provided below</p></br>
Portions © 1987-1989 Dow Jones & Company, Inc., © 2002 BBNT Solutions, LLC, © 1995, 1999, 2005, 2008, 2009 Trustees of the University of Pennsylvania

<h3>简介</h3><br><p>2008年CoNLL共享任务数据集由语言数据联盟（Linguistic Data Consortium, 缩写LDC）发布，目录号为LDC2009T12，ISBN为1-58563-505-7，包含<a href="http://www.conll.org/previous-tasks" rel="nofollow">2008年计算自然语言学习会议（Conference on Computational Natural Language Learning, CoNLL）共享任务评估</a>所用的试用语料库、训练语料库、开发集与测试集。2008年共享任务制定了句法依存标注规范，涵盖命名实体边界信息，以及动词谓词与名词谓词的语义依存关系模型标注。本次共享任务的语料素材节选自以下语料库：<a href="http://catalog.ldc.upenn.edu/LDC99T42" rel="nofollow">树库3（Treebank-3）</a> LDC99T42、<a href="http://catalog.ldc.upenn.edu/LDC2005T33" rel="nofollow">BBN代词共指与实体类型语料库</a> LDC2005T33、<a href="http://catalog.ldc.upenn.edu/LDC2004T14" rel="nofollow">命题库I（Proposition Bank I）</a> LDC2004T14（简称PropBank），以及<a href="http://catalog.ldc.upenn.edu/LDC2008T23" rel="nofollow">NomBank v1.0</a> LDC2008T23。</p><br><p><a href="http://www.conll.org/previous-tasks" rel="nofollow">计算自然语言学习会议（Conference on Computational Natural Language Learning, CoNLL）</a>每年都会配套举办共享任务，旨在推动自然语言处理应用的发展，并在标准化环境中对其进行评估。2004年与2005年的CoNLL共享任务专注于单语环境（英语）下的语义角色标注（Semantic Role Labeling, SRL）。2006年与2007年的共享任务则聚焦于句法依存句法分析，所用语料涵盖多达13种语言。2008年的共享任务采用统一的基于依存的形式化体系，将句法依存句法分析任务与语义论元识别及语义角色标注任务进行了融合。</p><br><p>LDC还发布了以下CoNLL共享任务数据集：</p><br><ul><br><li>2006年CoNLL共享任务——十种语言（<a href="../../../LDC2015T11">LDC2015T11</a>）</li><br><li>2006年CoNLL共享任务——阿拉伯语与捷克语（<a href="../../../LDC2015T12">LDC2015T12</a>）</li><br><li>2009年CoNLL共享任务 第一部分（<a href="../../../LDC2012T03">LDC2012T03</a>）</li><br><li>2009年CoNLL共享任务 第二部分（<a href="../../../LDC2012T04">LDC2012T04</a>）</li><br><li>2015-2016年CoNLL共享任务（<a href="../../../LDC2017T13">LDC2017T13</a>）</li><br></ul><br><h3>数据集</h3><br><p>2008年共享任务分为三个子任务：</p><br><ol><br><li>句法依存句法分析</li><br><li>语义谓词的识别与消歧</li><br><li>论元识别及针对每个谓词的语义角色赋值</li><br></ol><br><p>本次共享任务达成了多项目标：</p><br><ul><br><li>针对句法与语义依存关系均采用基于依存的表征方式来完成语义角色标注并进行评估。尽管基于依存树库的语义角色标注此前已有研究，但2008年共享任务的方法具有以下创新点：<br><ol><br><li>成分结构到依存结构的转换策略覆盖了命题库（PropBank）与NomBank v1.0中所有标注的语义论元，而非仅部分子集；</li><br><li>本次标注同时针对动词谓词（对应PropBank）与名词谓词（对应NomBank）构建了以命题为中心的标注体系。</li><br></ol><br></li><br><li>基于“更丰富的句法依存集合可改善语义处理效果”这一观察，本次任务所建模的句法依存关系相较于此前CoNLL共享任务所用的更为复杂。例如，本次语料库包含同位关系链接、源自命名实体（Named Entity, NE）结构的依存关系，以及对长距离语法关系的更优建模。</li><br><li>为句法与语义依存关系的联合学习提供了实用框架。</li><br></ul><br><p>由于2008年共享任务的复杂度较高，本次任务仅采用了英语单语言数据集。</p><br><h3>示例</h3><br><p>下文提供了一个共享任务标注的示例：</p><br><p>本数据集部分内容 © 1987-1989 Dow Jones & Company, Inc.、© 2002 BBNT Solutions, LLC、© 1995、1999、2005、2008、2009 宾夕法尼亚大学校董会</p>

搜集汇总

背景与挑战

背景概述

该数据集是2008年CoNLL（计算自然语言学习会议）共享任务的评估数据，专门用于英语自然语言处理研究。它包含从新闻通讯社和新闻杂志中提取的语料，重点支持句法依存分析和语义角色标注的联合任务，涉及动词和名词谓词的语义依存关系建模，旨在促进更复杂的语言处理框架开发。