Compound embedding features for semi-supervised learning

To solve data sparsity problem, recently there has been a trend in discriminative methods of NLP to use representations of lexical items learned from unlabeled data as features. In this paper, we investigated the usage of word representations learned by neural language models, i.e. word embeddings. The direct usage has disadvantages such as large amount of computation, inadequacy with dealing word ambiguity and rare-words, and the problem of linear non-separability. To overcome these problems, we instead built compound features from continuous word embeddings based on clustering. Experiments showed that the compound features not only improved the performances on several NLP tasks, but also ran faster, suggesting the potential of embeddings.