Learning Distributed Word Representations and Applications in Biomedial Natural Language Processing

A common challenge for biomedical natural language processing (BioNLP) is data sparsity. Distributed word representation approaches have been developed recently that represent word by learning from a large unlabeled data, such approaches have shown to overcome the data sparsity and improve NLP tasks. In this tutorial, we will introduce various word representation approaches, including the traditional Brown Clusters and the more recent distributed word representation (word embeddings), and their potential to improve other NLP tasks. Although we primarily discuss BioNLP applications, the methodologies are applicable to most of learning-based approaches (e.g., image processing and machine translation).