Unsupervised Learning of Style-sensitive Word Vectors
This work addresses the need for style-sensitive word vectors in natural language processing, representing an incremental advancement in unsupervised learning techniques.
The paper tackles the problem of capturing stylistic similarity between words in an unsupervised manner by extending the CBOW model with a wider context window, and it demonstrates that this approach contributes to learning style-sensitive word embeddings through experiments on a newly created benchmark dataset.
This paper presents the first study aimed at capturing stylistic similarity between words in an unsupervised manner. We propose extending the continuous bag of words (CBOW) model (Mikolov et al., 2013) to learn style-sensitive word vectors using a wider context window under the assumption that the style of all the words in an utterance is consistent. In addition, we introduce a novel task to predict lexical stylistic similarity and to create a benchmark dataset for this task. Our experiment with this dataset supports our assumption and demonstrates that the proposed extensions contribute to the acquisition of style-sensitive word embeddings.