Viet-Anh on Software Logo

What is: fastText?

SourceEnriching Word Vectors with Subword Information
Year2000
Data SourceCC BY-SA - https://paperswithcode.com

fastText embeddings exploit subword information to construct word embeddings. Representations are learnt of character nn-grams, and words represented as the sum of the nn-gram vectors. This extends the word2vec type models with subword information. This helps the embeddings understand suffixes and prefixes. Once a word is represented using character nn-grams, a skipgram model is trained to learn the embeddings.