Multi-View Learning of Word Embeddings via CCA
Multi-View Learning of Word Embeddings via CCA
复制标题
DOI：
--
复制
发表时间：
2011-12
期刊：
影响因子：
--
通讯作者：
Paramveer S. Dhillon;Dean P. Foster;L. Ungar
中科院分区：
文献类型：
--
作者：
Paramveer S. Dhillon;Dean P. Foster;L. Ungar
最近，人们对使用大量未标记的数据来学习单词表示产生了浓厚的兴趣，然后可以将其用作NLP任务的监督分类器中的特征。然而，目前大多数方法训练速度慢，没有对单词的上下文进行建模，并且缺乏理论基础。在本文中，我们提出了一种新的学习方法，低秩多视图学习（LR-MVL），它使用快速谱方法来估计低维上下文特定的词表示从未标记的数据。然后，这些表示特征可以与任何监督学习器一起使用。LR-MVL非常快，保证收敛到全局最优值，理论上很优雅，并且在命名实体识别（NER）和组块问题上实现了最先进的性能。
Recently, there has been substantial interest in using large amounts of unlabeled data to learn word representations which can then be used as features in supervised classifiers for NLP tasks. However, most current approaches are slow to train, do not model the context of the word, and lack theoretical grounding. In this paper, we present a new learning method, Low Rank Multi-View Learning (LR-MVL) which uses a fast spectral method to estimate low dimensional context-specific word representations from unlabeled data. These representation features can then be used with any supervised learner. LR-MVL is extremely fast, gives guaranteed convergence to a global optimum, is theoretically elegant, and achieves state-of-the-art performance on named entity recognition (NER) and chunking problems.