Using Centroids of Word Embeddings and Word Mover's Distance for Biomedical Document Retrieval in Question Answering

摘要

We propose a document retrieval method for question answering that representsdocuments and questions as weighted centroids of word embeddings and reranksthe retrieved documents with a relaxation of Word Mover's Distance. Usingbiomedical questions and documents from BIOASQ, we show that our method iscompetitive with PUBMED. With a top-k approximation, our method is fast, andeasily portable to other domains and languages.