Sentence Embedding

An unsupervised algorithm for textual similarity tasks. Uses word embeddings to represent sentence by the weighted average of word vectors and then modified by singular valued decomposition.