Jointly Learning Word Representations and Composition Functions Using Predicate-Argument Structures
Jointly Learning Word Representations and Composition Functions Using Predicate-Argument Structures
复制标题
DOI：
10.3115/v1/d14-1163
复制
发表时间：
2014-10
期刊：
影响因子：
--
通讯作者：
Kazuma Hashimoto;Pontus Stenetorp;Makoto Miwa;Yoshimasa Tsuruoka
中科院分区：
文献类型：
--
作者：
Kazuma Hashimoto;Pontus Stenetorp;Makoto Miwa;Yoshimasa Tsuruoka
我们介绍了一种新的构图语言模型，该模型工作在预测参数结构(PASS)上。我们的模型使用bagof-word和基于依存关系的上下文联合学习单词表示及其组成功能。与以前的基于词序的模型不同，我们的基于PAS的模型通过使用PAS中的类别信息将论元组成谓词。这使我们的模型能够捕获单词之间的长期依存关系，并更好地处理动词宾语和主语-动词-宾语关系等结构。我们使用两个短语相似度数据集进行了实验验证，得到了与之前最好的结果相当或更高的结果。我们的系统在不需要预先训练的词向量和使用小得多的训练语料库的情况下获得了这些结果；尽管如此，对于主语-动词-宾语数据集，我们的模型在相对性能上比现有技术提高了高达∼10%。
We introduce a novel compositional language model that works on PredicateArgument Structures (PASs). Our model jointly learns word representations and their composition functions using bagof-words and dependency-based contexts. Unlike previous word-sequencebased models, our PAS-based model composes arguments into predicates by using the category information from the PAS. This enables our model to capture longrange dependencies between words and to better handle constructs such as verbobject and subject-verb-object relations. We verify this experimentally using two phrase similarity datasets and achieve results comparable to or higher than the previous best results. Our system achieves these results without the need for pretrained word vectors and using a much smaller training corpus; despite this, for the subject-verb-object dataset our model improves upon the state of the art by as much as ∼10% in relative performance.